LogoFDA 510(k) Search
K Number
DEN130011
Device Name
ILLUMINA MISEQDX PLATFORM
Manufacturer
ILLUMINA, INC.
Date Cleared
2013-11-19

(57 days)

Product Code
PFF
Regulation Number
862.2265
Type
Post-NSE
Panel
Clinical Chemistry
Reference & Predicate Devices
k133136
Predicate For
N/A
AI/MLSaMDIVD (In Vitro Diagnostic)TherapeuticDiagnosticis PCCP AuthorizedThirdpartyExpeditedreview
Intended Use

The MiSeqDx Platform is a sequencing instrument that measures fluorescence signals of labeled nucleotides through the use of instrument specific reagents and flow cells (MiSeqDx Universal Kit 1.0), imaging hardware, and data analysis software. The MiSeqDx Platform is intended for targeted sequencing of human genomic DNA from peripheral whole blood samples. The MiSeqDx Platform is not intended for whole genome or de novo sequencing.

Device Description

The MiSeqDx Platform is a high throughput DNA sequence analyzer for clinical use.

The MiSeqDx Platform consists of the MiSeqDx instrument and data analysis software. It is for use with the MiSeqDx Universal Kit 1.0 [MiSeqDx reagent cartridge, MiSeqDx flow cell, SBS Solution (PR2 buffer)] for library preparation and sample indexing (K133136). The end-user inputs extracted genomic DNA to be sequenced and provides the Analyte Specific Reagents (ASRs) to develop a sequencing assay that targets their sequence of interest.

AI/ML Overview

The Illumina MiSeqDx Platform is a high-throughput DNA sequencing instrument. The acceptance criteria and supporting studies are detailed below:

1. Table of Acceptance Criteria and Reported Device Performance

The acceptance criteria for the MiSeqDx Platform are primarily derived from the special controls stipulated in 21 CFR 862.2265, focusing on accuracy and reproducibility across various genomic features.

Acceptance Criteria (from 21 CFR 862.2265 (B) items ii, iii, vii, viii)Reported Device Performance (from Accuracy and Reproducibility Studies)
Accuracy:
- Ability to detect single nucleotide variants (SNVs).Study 1: All SNVs had 100% agreement with the reference sequence. PPA ranged from 89.5% to 95.8% (due to missed indels, not SNVs). NPA was 100%. Study 2: PPA for SNVs and Indels was 94.1%, NPA was 100%.
- Ability to detect insertions and deletions (indels).Study 1: Variants missed were 1-base insertions or 1-base deletions in homopolymer regions (e.g., Amplicon 9 and 95). Study 3 (CFTR Assay): 1-base insertion, 3-base deletion, and 2-base deletion were detected with 100% correct calls. Overall: Validated for detection of SNVs and up to 3-base deletions. Evaluation of 1-base insertions was limited to 3 different insertions on 3 separate chromosomes. The system has problems detecting 1-base insertions or deletions in homopolymer tracts. 2 out of 3 1-base insertions tested were called correctly (those in non-homopolymer regions). 3 out of 4 1-base deletions called correctly (those in non-homopolymer regions).
- Performance across varying sequence context (e.g., GC-rich regions, homopolymer runs, different chromosomes).Study 1 & 3 summary: - GC content > 19% and < 72%: 100% correct calls for all bases in 135 out of 135 sequenced amplicons within these ranges. - PolyA lengths ≤ 7: 100% correct calls in 270 out of 270 amplicons. - PolyT lengths ≤ 8: 100% correct calls in 270 out of 270 amplicons. - PolyG lengths ≤ 6: 100% correct calls in 405 out of 405 amplicons. - PolyC lengths ≤ 7: 100% correct calls in 135 out of 135 amplicons. - Dinucleotide repeat lengths < 5x: 100% correct calls in 135 out of 135 amplicons. - Trinucleotide repeat lengths ≤ 4x: 100% correct calls in 810 out of 810 amplicons. - Specific limitations noted for homopolymer runs exceeding eight bases (filtered out in VCF) and problems with 1-base insertions/deletions in homopolymer tracts.
Reproducibility:
- Consistency across multiple instruments, operators, and sites.Study 1 (General Amplicon Panel): Variants were reproducible across nine runs for the samples tested, showing identical results for replicate samples. Incorrect calls and no calls for certain amplicons were consistently observed across all three MiSeqDx instruments, often linked to homopolymer regions. Study 2 (CFTR Assay): Overall agreement typically 100% across 3 sites and 2 operators per site. A few cases had lower agreement (e.g., Sample Y122X/R1158X (HET) and F508del (HET) with ~94.4% overall agreement due to no calls or miscalls, often isolated to one site/operator). For variants, Positive Agreement was mostly 100%, with occasional drops to 97.22% or 94.44% due to miscalls or no calls. Negative Agreement was consistently 100% or 99.96%.
Limitations:
- Specification of sequence variations not detectable with claimed accuracy/precision.Variants in homopolymer runs exceeding eight bases will be filtered out (R8 filter). The system has problems detecting 1-base insertions or deletions in homopolymer tracts. Evaluation of 1-base insertions was limited to 3 different insertions.
Interfering Substances:Interfering Substances Study: 100% call rate for samples tested with bilirubin, cholesterol, hemoglobin, triglycerides, and EDTA at specified concentrations.
DNA Input Range:DNA Input Study: 100% accuracy and call rate for DNA inputs between 25 ng and 1250 ng. Two no calls observed at 25ng, resulting in 99.26% sample call rate for those samples.
Other Supportive Data:
- DNA Extraction Methods.DNA Extraction Study: Alcohol precipitation, silica filter column isolation, and magnetic bead extraction methods all yielded 100% call rate, 100% accuracy, and 100% sample first pass rate.
- Thermal Cycler Study.Thermal Cycler Study: Met sponsor's acceptance criteria, demonstrating commercial thermal cyclers are adequate.
- Sample Indexing.Sample Indexing Study: 100% reproducibility and accuracy for all sample/index primer combinations across 96 different index primers.
- Specimen Storage and Freeze-Thaw.Specimen Storage/Freeze-Thaw Studies: No miscalls or no calls observed for any specimens, demonstrating tested blood and gDNA storage conditions did not affect assay results.

2. Sample Sizes Used for the Test Set and Data Provenance

  • Accuracy Study 1:
    • Sample Size: 13 unique human genomic DNA samples (from two parents and 11 children). These 13 samples were run in 15 instances (two samples in duplicate). The study queried 24,434 bases across 19 different chromosomes.
    • Data Provenance: The origin of the samples is implied to be human, likely from a well-characterized cohort given the reference to "frequently sequenced by multiple laboratories and sequencing methodologies". The study itself is retrospective/analytical performance data. No specific country of origin is mentioned.
  • Accuracy Study 2:
    • Sample Size: 1 sample (NA12878). This sample included 184 amplicons within highly confident reference calls.
    • Data Provenance: The sample (NA12878) is a well-known reference sample established by the National Institutes of Standards and Technology (NIST). This is retrospective data, comparing the device's output to an established reference. No specific country of origin is mentioned.
  • Accuracy Study 3:
    • Sample Size: Six samples (implicitly, since the table shows 6 amplicons with 1 "sample" column). The study evaluated "a subset of CFTR clinically significant genetic variations".
    • Data Provenance: These samples were "characterized by bidirectional Sanger sequencing as a reference method." This is retrospective analytical performance data. No specific country of origin is mentioned.
  • Reproducibility Study 1:
    • Sample Size: 13 unique human genomic DNA samples similarly to Accuracy Study 1. These samples were run over nine runs, with each run generating results for 15 samples (total of 135 samples run for each amplicon). For lot-to-lot reproducibility, 94 samples and two non-template controls were tested across three lots.
    • Data Provenance: Same as Accuracy Study 1, human samples, retrospective analytical performance data.
  • Reproducibility Study 2:
    • Sample Size: Two well-characterized panels of 46 samples each were tested. This resulted in a total of 810 calls per site (46 samples * 2 operators * 3 sites * X amplicons, though the table shows 810 total calls per specific genotype).
    • Data Provenance: Genomic DNA from cell lines with known variants and leukocyte-depleted blood spiked with cell lines with known variants in the CFTR gene. This is retrospective analytical performance data.
  • Interfering Substances Study:
    • Sample Size: Eight whole blood samples representing eight unique genotypes. 48 replicates for each interfering substance test.
  • DNA Extraction Study:
    • Sample Size: 14 unique blood samples. Total sample size for each extraction method was 168 (14 samples x 2 operators/extraction method x 3 runs/operator x 2 replicates/extracted gDNA sample).
  • DNA Input Study:
    • Sample Size: 14 representative DNA samples. Tested in duplicate at 9 DNA input levels. For individual input levels (1250 ng, 250 ng, 100 ng), 4 samples x 20 replicates (80 samples). For 25 ng, 14 samples x 20 replicates (280 samples).
  • Sample Indexing Study:
    • Sample Size: 8 unique DNA samples tested with 96 different indexing primer combinations.
  • Specimen Storage/Freeze-Thaw Studies:
    • Sample Size: Six K2EDTA anti-coagulated blood samples (divided into 6 aliquots each) for storage. 15 DNA samples for freeze-thaw.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications

The documentation does not explicitly state the "number of experts" or their specific qualifications (e.g., "radiologist with 10 years of experience") for establishing ground truth. Instead, it refers to:

  • "Well-characterized composite reference information": This reference database was "derived from the combination of multiple sequencing methodologies, publicly available data, and hereditary information." This implies a consensus approach from various robust sources rather than individual expert adjudication.
  • Highly confident genotype established for NA12878 by the National Institutes of Standards and Technology (NIST): NIST reference materials are developed through extensive, rigorous characterization by specialized scientific teams, often involving multiple technologies and consensus building.
  • "Bidirectional Sanger sequencing as a reference method": Sanger sequencing is a gold standard for sequence verification. The interpretation of Sanger data typically involves trained personnel (e.g., geneticists, molecular biologists).
  • For the reproducibility study, the reference was also "well characterized reference database."

4. Adjudication Method for the Test Set

The concept of "adjudication method" (like 2+1 or 3+1) is typically associated with human reviewer discrepancies, often in image-based diagnostics. For genomic sequencing, ground truth is established using orthogonal, highly accurate sequencing technologies and/or composite reference databases. Therefore, traditional "adjudication" by multiple human experts in the sense of reconciling differing interpretations is not directly applicable here. The ground truth generation methods inherently build in a form of "adjudication" or consensus through:

  • Composite Reference Information: Combining data from multiple sequencing methods, public databases, and hereditary information.
  • NIST Highly Confident Genotype: A highly rigorous, multi-method approach by a standards organization.
  • Bidirectional Sanger Sequencing: Often considered a definitive (gold standard) method, interpretation relies on the inherent reliability of the technology.

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study

No, a Multi-Reader Multi-Case (MRMC) comparative effectiveness study was not done. MRMC studies are typically performed in diagnostic imaging to assess the impact of AI on human reader performance. The MiSeqDx Platform is an instrument for high-throughput DNA sequencing, and its performance is evaluated directly against established genomic references, not through human interpretation of complex data that could be aided by AI.

6. Standalone (Algorithm Only Without Human-in-the-Loop) Performance

Yes, the studies entirely describe the standalone performance of the MiSeqDx Platform (instrument and its integrated software for data analysis, base calling, and variant calling). The performance metrics (accuracy, reproducibility, call rate) are reported for the system operating without human interpretation influencing the primary result generation. Human intervention is limited to sample preparation and initiating runs, not in the sequence interpretation itself by the device. The MiSeqDx Reporter software performs de-multiplexing and FASTQ file generation "without user intervention".

7. Type of Ground Truth Used

The ground truth used was primarily:

  • Well-characterized composite reference information: This involved combining data from multiple sequencing methodologies, publicly available data, and hereditary information.
  • Highly confident genotype established by the National Institutes of Standards and Technology (NIST): A gold standard reference.
  • Bidirectional Sanger sequencing data: Another gold standard for sequence verification.
  • PCR assay (for deletions): A molecular biology technique used to confirm the presence or absence of specific sequences.
  • Human genome reference sequence build 19 (GRCh37/hg19): Used for non-variant base comparisons.

8. Sample Size for the Training Set

The document does not explicitly mention a "training set" in the context of machine learning or AI algorithms. The MiSeqDx is a sequencing platform, and its "software" (RTA and MiSeqDx Reporter) performs base calling, de-multiplexing, sequence alignment, and variant calling. If there were any machine learning components in these software modules, the training data used to develop them are not detailed in this document. The studies described are validation studies of the finalized device's performance.

9. How the Ground Truth for the Training Set Was Established

As no specific "training set" for an AI or machine learning algorithm is identified or discussed, the method for establishing its ground truth is not provided. The described ground truth methods (composite reference, NIST, Sanger) are for the test/validation sets used to assess the final device performance.

{0}------------------------------------------------

EVALUATION OF AUTOMATIC CLASS III DESIGNATION FOR MISEQDX PLATFORM

DECISION SUMMARY INSTRUMENT ONLY TEMPLATE Correction Date: February 24, 2017

This Decision Summary contains corrections to the November 19, 2013 Decision Summary.

A. 510(k) Number:K123989
B. Purpose for Submission:De novo request for evaluation of automatic class III designation for the Illumina MiSeqDx Platform
C. Type of Test or Tests Performed:High-throughput DNA sequencing
D. Applicant:Illumina Inc.
E. Device Name:MiSeqDx Platform

F. Regulatory Information:

FDA identifies this type of device as:

    1. New regulation number: 21 CFR 862.2265
    1. Classification: Class II.
    1. Product code: PFF High throughput DNA sequence analyzer
    1. Panel: Toxicology (91)

G. Intended Use:

  • 1.Intended uses(s):
    The MiSeqDx Platform is a sequencing instrument that measures fluorescence signals of labeled nucleotides through the use of instrument specific reagents and flow cells (MiSeqDx Universal Kit 1.0), imaging hardware, and data analysis software. The MiSeqDx Platform is intended for targeted sequencing of human genomic DNA from peripheral whole blood samples. The MiSeqDx Platform is not intended for whole genome or de novo sequencing.

    1. Indication for uses(s):
      Same as intended use above.

{1}------------------------------------------------

3. Special conditions for use statement(s):

    1. This product is limited to delivering:
    • Sequencing output >1 Gb o
    • Reads >3 million o
    • o Read length (in paired end run) 2 x 150 bp
    • Bases higher than 030 >75% (Greater than 75% of bases have Phred scale . quality score greater than 30, indicating base call accuracy greater than 99.9%)
    1. Variants in homopolymer runs exceeding eight bases will be filtered out in the VCF files (R8 filter).
  • The system has been validated for the detection of SNVs and up to 3 base deletions. 3. Evaluation of 1 base insertions was been limited to 3 different insertions on 3 separate chromosomes.
  • The system has problems detecting 1 base insertions or deletions in homopolymer 4. tracts (e.g., polyA).
    1. This MiSeqDx system is designed to deliver qualitative (i.e. genotype) results.
    1. As with any hybridization-based workflow, underlying polymorphisms or mutations in oligonucleotide-binding regions can affect the alleles being probed and, consequently, the calls made.
    1. Recommended minimal coverage per amplicon needed for accurate variant calling (Q(max gt | poly site) >= 100) is 75x.

H. System Descriptions:

1. Device Description:

The MiSeqDx Platform is a high throughput DNA sequence analyzer for clinical use.

The MiSeqDx Platform consists of the MiSeqDx instrument and data analysis software. It is for use with the MiSeqDx Universal Kit 1.0 [MiSeqDx reagent cartridge, MiSeqDx flow cell, SBS Solution (PR2 buffer)] for library preparation and sample indexing (K133136). The end-user inputs extracted genomic DNA to be sequenced and provides the Analyte Specific Reagents (ASRs) to develop a sequencing assay that targets their sequence of interest.

2. Principles of Operation:

Testing begins with genomic DNA extracted from a peripheral whole blood sample. The genomic DNA is processed through library preparation. which specifically amplifies the intended genomic regions of each sample while also adding the indexes and flow cell capture sequences to the amplified products. The resulting sample libraries are then transferred into a MiSeqDx reagent cartridge which contains all of the reagents required for cluster generation and sequencing (Sequencing By Synthesis - SBS). The MiSeqDx cartridge, MiSeqDx flow cell, and MiSeqDx SBS Solution (PR2 buffer) are then inserted

{2}------------------------------------------------

into the MiSeqDx instrument which performs cluster generation, sequencing and data analysis.

The instrument uses cluster generation on the flow cell surface followed by sequencing using the Sequencing by Synthesis (SBS) process.

After the flow cell images are captured by the MiSeqDx instrument following each sequencing cycle, primary analysis is performed without user intervention. Primary analysis is performed by the RTA (Real Time Analysis) software, and consists of base calling of each cluster at each cycle. In addition to calling the bases, RTA assigns an analytical quality score (O-score) to each base call. Calculations of O-scores are based on the ratio of the signal intensity of the highest base in a given cluster during a given cycle to the signal intensity of the three other bases. The quality score Q is calculated as -10 log10 P, where P is the probability that base call is incorrect.

Secondary analysis is performed by the MiSeqDx Reporter software. It also occurs without user intervention and consists of de-multiplexing and FASTO file generation. De-multiplexing is the process of using the index sequences to assign clusters to the sample from which they originated.

After base calling and de-multiplexing, the software generates FASTQ files that contain sequence and quality information. Due to the massively parallel nature of the SBS biochemistry, hundreds of independent sequencing reads, each with their own quality score, are generated for each amplicon in each sample. The FASTO file which is a widely accepted text based format for storing both a nucleotide sequence and its corresponding quality score. FASTQ files serve as input files for various sequence alignment and subsequent variant calling algorithms.

The MiSeqDx has a sequence alignment and variant calling program available for use.

    1. Modes of Operation:
      The MiSeqDx is a high throughput nucleic acid analyzer.

4. Specimen Identification:

Samples up to 96 unique specimens can be analyzed. Eight unique index primer sequences (forward), named i5 primers, and 12 unique index primer (reverse) sequences, named i7 primers, are provided. These 8 unique forward index primers and 12 unique reverse index primers, when combined in a pair wise manner, produce 96 unique index combinations allowing for up to 96 samples to be processed in parallel during the library preparation process. These are added during the library preparation process. The sample sheet, a file that the user provides the software, contains the link between each of the sample names and their associated index sequences.

{3}------------------------------------------------

After completion of the sequence run, MiSeq Reporter software de-multiplexes the samples using the index sequences and creates FASTQ files as the data analysis output. The user can also utilize the MiSeq Reporter Software for sequence alignment and variant calling.

5. Specimen Sampling and Handling:

The MiSeqDx specimen is a pooled library (or libraries) derived from genomic DNA extracted from peripheral whole blood that then undergoes the following steps to create the pooled library: the genomic DNA is quantified and then used to make a library, the library sample is processed to remove remaining library preparation reagents (e.g. unused primers), normalized, and then pooled for input on the analyzer. Library normalization is used to ensure that each library is equally represented in the pooled sample.

At a minimum, eight samples must be present. If six unique samples (excluding the positive and negative controls) are not available, it is acceptable to fill the run with sample replicates or any human genomic DNA sample.

    1. Calibration:
      There is no end-user calibration of the system. During installation of the platform, a company representative (Field Applications Scientist) begins a series of tests to validate the performance of the instrument subsystems, which include optical alignment, fluidic delivery, and thermal calibration, among others. In the case of a test failure, the MiSeqDx company representative uses a set of instrument-specific tools to adjust and/or repair the instrument to meet operational specifications. Re-calibration occurs during the preventive maintenance visit.
    1. Quality Control:
      A PhiX internal control (i.e. genomic DNA from the bacteriophage ΦΧ174) is added to each pooled library prior to placement on the instrument. Successful sequencing of the PhiX genome indicates that the sequencing chemistry worked as expected. A negative control, or no template control, (not provided by the sponsor) should be included in every run in order to detect the presence of contamination in the environment or run.
    1. Software:
      FDA has reviewed applicant's Hazard Analysis and Software Development processes for this line of product types:

Yes X

{4}------------------------------------------------

H. Substantial Equivalence Information:

    1. Predicate Device Name(s) and 510(k) numbers:
      Not applicable.
    1. Comparison with Predicate Device:
      Not applicable.

I. Special Control/Guidance Document Referenced (if applicable):

Not applicable.

J. Performance Characteristics:

    1. Analytical Performance:
    • a. Accuracy:

Three accuracy studies were conducted.

Study 1: This accuracy study used a representative assay designed to query a variety of genes covering 24,434 bases across 19 different chromosomes, and containing potentially clinically relevant exons. The 13 unique samples used in this study are from two parents and 11 children that have been frequently sequenced by multiple laboratories and sequencing methodologies. There are six samples from females and seven from males.

Accuracy was determined for single nucleotide variants (SNVs) by comparing the study data to well-characterized composite reference information. The reference database sequence was derived from the combination of multiple sequencing methodologies, publicly available data, and hereditary information. The following table to evaluate accuracy of the system was compiled based on data from the first run in the study. No repeat testing was done for this study.

Amp-liconChr.Ana-lyzedfrag-mentsize1AmpliconGenomicContent# ofuniquesamplestotal # ofsamplesanalyzed2#calls/samplethatcouldbemade3# ofnocalls4# ofcorrectcalls/sample5# in-correctcalls6%correctcalls7
11132Poly C (5),63% GC131513201320100.00
21128Poly T (5)131512801280100.00
32133-131513301330100.00
Amp-liconChr.Ana-lyzedfrag-mentsize¹AmpliconGenomicContent# ofuniquesamplestotal # ofsamplesanalyzed²#calls/samplethatcouldbemade³# ofnocalls⁴# ofcorrectcalls/sample⁵# in-correctcalls⁶%correctcalls⁷
42119-131511901190100.00
52127Poly T (5)131512701270100.00
62135Poly A (6)131513501350100.00
72122Poly T (5),Poly C (5)131512201220100.00
82110Poly T (5)131511001100100.00
9⁸2131Poly A (14)1315130-1310130-131999.54
102117-131511701170100.00
112121-131512101210100.00
122114-131511401140100.00
132129Poly A (5)131512901290100.00
143131Poly A (5),Poly T (5)131513101310100.00
153130-131513001300100.00
163130-131513001300100.00
173117-131511701170100.00
183136Poly T (5)131513601360100.00
193131Poly T (5),SNV131513101310100.00
203123Poly A (5)131512301230100.00
213117Poly A (6),Poly T (5),Homologousregion on adifferentchromosome131511701170100.00
223119Homologousregion on adifferentchromosome131511901190100.00
233120-131512001200100.00
243129Poly T (5)131512901290100.00
254133Poly C (7),66% GC131513301330100.00
264135Poly C (5),60% GC131513501350100.00
Amp-liconChr.Analyzedfragmentsize1AmpliconGenomicContent# ofuniquesamplestotal # ofsamplesanalyzed2# calls/sample thatcould bemade3# ofnocalls4# ofcorrectcalls/sample5# in-correctcalls6%correctcalls7
274123SNV131512301230100.00
284134-131513401340100.00
294132-131513201320100.00
304121Poly A (5),SNV131512101210100.00
314125-131512501250100.00
324134Poly T (5)131513401340100.00
334118-131511801180100.00
344122Poly A (5)131512201220100.00
354131-131513101310100.00
364133-131513301330100.00
374128Poly T (6)131512801280100.00
384131-131513101310100.00
394129Poly A (5),Poly T (5),SNV131512901290100.00
404133Poly T (5),SNV131513301330100.00
414112SNV131511201120100.00
424133-131513301330100.00
434135-131513501350100.00
444122-131512201220100.00
454117-131511701170100.00
464124-131512501250100.00
474117Poly T (5)131511701170100.00
484128Poly A (7)131512801280100.00
494123Poly A (6)131512301230100.00
504133-131513301330100.00
514112-131511201120100.00
524129-131512901290100.00
534126-131512601260100.00
544132-131513201320100.00
555131-131513101310100.00
565119-131511901190100.00
575120Poly A (5)131512001200100.00
585119-131511901190100.00
Amp-liconChr.Ana-lyzedfrag-mentsize1AmpliconGenomicContent# ofuniquesamplestotal # ofsamplesanalyzed2#calls/samplethatcouldbemade3# ofnocalls4# ofcorrectcalls/sample5# in-correctcalls6%correctcalls7
595118-131511801180100.00
605112-131511201120100.00
615120-131512001200100.00
625120Poly A (5)131512001200100.00
635115CT(5)131511501150100.00
645112SNV131511201120100.00
655135Poly T (6)131513501350100.00
66513163% GC131513101310100.00
675121-131512101210100.00
685132Poly A (6),Poly T (8)131513201320100.00
697133-131513301330100.00
70712060% GC131512001200100.00
717135-131513501350100.00
727126Poly A (5),59% GC131512601260100.00
737134-131513401340100.00
747122Poly C (5),63% GC131512201220100.00
75712759% GC;SNV131512701270100.00
767123-131512301230100.00
777125-131512501250100.00
787133Poly A (5),Poly T (5)131513301330100.00
797116-131511601160100.00
807135-131513501350100.00
817118-131511801180100.00
82713667% GC131513601360100.00
83713158% GC131513101310100.00
847119Poly G (6),61% GC131511901190100.00
857122Poly T (5)131512201220100.00
867123Poly A (6)131512301230100.00
87812760% GC131512701270100.00
88812957% GC131512901290100.00
Amp-liconChr.Ana-lyzedfrag-mentsize¹AmpliconGenomicContent# ofuniquesamplestotal # ofsamplesanalyzed²#calls/samplethatcouldbemade³# ofnocalls4# ofcorrectcalls/sample5# in-correctcalls6%correctcalls7
899130Poly T (5)131513001300100.00
909116-131511601160100.00
919119Homologousregion on adifferentchromosome131511901190100.00
929121-131512101210100.00
939117Homologousregion on adifferentchromosome131511701170100.00
949114-131511401140100.00
95109129Poly A (14)13151300129 (of130)1599.23
969114Homologousregion on adifferentchromosome;SNV131511401140100.00
979122-131512201220100.00
989127Poly A (5),Poly C (5)131512701270100.00
999133-131513301330100.00
100913864% GC131513801380100.00
1019139-131513901390100.00
1029116-131511601160100.00
1039133Poly A (5),57% GC131513301330100.00
104913857% GC131513801380100.00
1059136Poly C (5),67% GC131513601360100.00
106911870% GC131511801180100.00
1071012862% GC131512801280100.00
1081012060% GC131512001200100.00
1091013958% GC;SNV131513901390100.00
1101011857% GC131511801180100.00

{5}------------------------------------------------

{6}------------------------------------------------

{7}------------------------------------------------

{8}------------------------------------------------

{9}------------------------------------------------

Amp-liconChr.Ana-lyzedfrag-mentsize'AmpliconGenomicContent# ofuniquesamplestotal # ofsamplesanalyzed2#calls/samplethatcouldbemade3# ofnocalls4# ofcorrectcalls/sample5# in-correctcalls6%correctcalls'
11110123Poly T (5)13ો ર12301230100.00
1121012113ા ર12101210100.00
1131012926% GC13ો ર12901290100.00
1141012213ા ર12201220100.00
ા । ર10124Poly T (5);Homologousregion on adifferentchromosome13ો ર12401240100.00
11610ા ૩૨CA(4)13ો રા ૩૨0ા રેર0100.00
11710ા ૩૨Poly A (6);Homologousregion on adifferentchromosome13ાં રા રેર0। ૩૨0100.00
11810119Poly C (5);SNV13ા ર11901190100.00
11910ા ટેર-13ો રા ટેર0ા ટેર0100.00
12010131-13ા ર13101310100.00
12110117-13ો ર11701170100.00
12210116-13ા ર11601160100.00
1231012958% GC13ો ર12901290100.00
12411117Poly T (10)13ા ર11701170100.00
। ટેર11117Poly T (5)13ો ર11701170100.00
12611113Poly A (5)13ા ર11301130100.00
12711129-13ા ર12901290100.00
12811121Poly T (5)13ો ર12101210100.00
12911123-13ો ર12301230100.00
13011127Poly A (6)13ા ર12701270100.00
13111136Poly T (6)13ા ર13601360100.00
13211132Poly T (5)13ો ર13201320100.00
13311ા । ર13ો રા । ર0ા । ર0100.00
13411117Poly T (8);19% GC13ા ર11701170100.00
ા રેર11134Poly A (5);Poly T (5)13ા ર13401340100.00
13611131Poly A (5)। ਤੇા ર13101310100.00
Amp-liconChr.Ana-lyzedfrag-mentsize¹AmpliconGenomicContent# ofuniquesamplestotal # ofsamplesanalyzed²#calls/samplethatcouldbemade³# ofnocalls4# ofcorrectcalls/sample5# in-correctcalls6%correctcalls7
1371113326% GC;SNV131513301330100.00
13811137Poly T (8);SNV131513701370100.00
13911131Poly A (5)131513101310100.00
14012131-131513101310100.00
14112128-131512801280100.00
14212133Poly A (5)131513301330100.00
14312136-131513601360100.00
14412124-131512401240100.00
1451212259% GC131512201220100.00
14613122-131512201220100.00
14713116Poly C (5)131511601160100.00
14813133-131513301330100.00
14913117SNV131511701170100.00
15013124Poly T (6)131512401240100.00
15113123Poly T (5);26% GC131512301230100.00
15213115Poly A (5)131511501150100.00
15313125-131512501250100.00
15413121-131512101210100.00
15513123-131512301230100.00
15613114-131511401140100.00
15713119-131511901190100.00
1581412258% GC131512201220100.00
15916122-131512201220100.00
16016121-131512101210100.00
16116123Poly C (5)131512301230100.00
16217119-131511901190100.00
1631711961% GC131511901190100.00
16417135-131513501350100.00
16517116Poly C (6);60% GC;SNV131511601160100.00
16617123-131512301230100.00
1671711662% GC131511601160100.00
Amp-liconChr.Ana-lyzedfrag-mentsize¹AmpliconGenomicContent# ofuniquesamplestotal # ofsamplesanalyzed²#calls/samplethatcouldbemade³# ofnocalls4# ofcorrectcalls/sample5# in-correctcalls6%correctcalls7
16817118Poly C (5);65% GC131511801180100.00
16917129-131512901290100.00
17017131Poly G (6);67% GC;SNV131513101310100.00
1711712761% GC131512701270100.00
17217118Poly C (5)131511801180100.00
1731713861% GC131513801380100.00
1741713158% GC131513101310100.00
17518112-131511201120100.00
17618124-131512401240100.00
17718134Poly A (6)131513401340100.00
17818129-131512901290100.00
17918133-131513301330100.00
18018118-131511801180100.00
1811811460% GC131511401140100.00
18218118-131511801180100.00
18319122Poly G (6);66% GC131512201220100.00
1841913964% GC131513901390100.00
1851913167% GC131513101310100.00
1861914159% GC;Homologousregion on adifferentchromosome131514101410100.00
18719121Poly C (5);72% GC;Homologousregion on adifferentchromosome131512101210100.00
1881913858% GC131513801380100.00
1891912364% GC131512301230100.00
19019138-131513801380100.00
19120117Poly T (5)131511701170100.00
Amp-liconChr.Ana-lyzedfrag-mentsize1AmpliconGenomicContent# ofuniquesamplestotal # ofsamplesanalyzed2#calls/samplethatcouldbemade3# ofnocalls4# ofcorrectcalls/sample5# in-correctcalls6%correctcalls7
19222136Poly A (7)131513601360100.00
19322122Poly A (5);Poly C (5)131512201220100.00
1942212262% GC;SNV131512201220100.00
1952211966% GC131511901190100.00

{10}------------------------------------------------

{11}------------------------------------------------

{12}------------------------------------------------

  1. Analyzed fragment means the size of the sequenced genomic region in bases, not including target-specific primers.

2 Total # of samples analyzed is 15 because two of the 13 unique samples were run in two independent replicates.

3 # calls/sample that could be made is the number of bases that had adequate quality to be called by the system

4 # of no calls is the number of bases in an amplicon that results in a no call in the run

5 # correct calls per sample is number of bases in the amplicon that were called that had results that matched the human genome reference sequence build 19+ and the well characterized composite reference.

6 # incorrect calls were the total number of incorrect calls for the SNV or indel in that amplicon; addition details on incorrect calls are presented in footnotes below.

7 % correct calls equals the correct call rate for all of the bases in the amplicon, where the correct call for the SNV or indel is based on the well characterized composite reference information and the correct call for the bases in the remainder of the amplicon sequence is based on comparison to human genome reference sequence build 19. This column may have more than one expected result for a given amplicon if some samples contain an indel while some do not, e.g., amplicon 9.

8 Amplicon 9 has a homopolymer run of 14 A's according to the human genome reference sequence build 19. However, the well characterized composite reference information for 7 out of 13 samples have 13 A 's in this homopolymer run. In these 7 samples, this one base pair deletion represents a false negative in the MiSeqDx sequencing accuracy study.

¹ Human Feb. 2009 (GRCh37/hg19) assembly available from NCBI

{13}------------------------------------------------

9 Amplicon 46 has a one base insertion which is reported in 9 samples in the well characterized reference database and is correctly detected in all analyzed samples.

10 Amplicon 95 has a homopolymer run of 14 A's according to human genome reference sequence build 19. However, the well characterized composite reference sequences for 13 out of 13 samples have 15 A 's in this homopolymer run. In these 13 samples, this one base pair insertion is a false negative in the MiSeqDx sequencing accuracy study.

The following table contains data from study 1 presented with positive and negative percent agreement (PPA and NPA, respectively), where the variant results are compared to the well characterized composite reference information for PPA calculations. Since the composite reference information only provides results for the single nucleotide variants and insertions/deletions. non-variant base results are compared to human genome reference sequence build 19 for NPA calculations. All non-variant bases had 100% agreement with the reference sequence. All SNVs had 100% agreement with the reference sequence. Variants that were missed were either 1 base insertions or 1 base deletions in the homopolymer regions.

Sample#amplicons%Amplicon1Coverage1Variantsexpectedpersample2VariantsCorrectlyCalledVariantsMissed3Non-variantbases calledcorrectlyPPA4NPA5
NA12877195100191722441889.5100
NA12878195100191722441789.5100
NA12879195100201912441695100
NA12880195100201822441790100
NA12881195100222022441590.9100
NA12882195100161512441993.8100
NA12883195100242312441295.8100
NA12884195100212012441595.2100
NA12885195100191722441789.5100
NA12886195100222022441590.9100
NA12887195100191812441694.7100
NA12888195100242312441295.8100
NA12893195100201822441790100

1 % Amplicon coverage is number of bases in the amplicons sequenced with confidence

2 Variants expected per sample includes both SNVs and indels

3 For the variants missed, please see the first table for study 1 and the footnotes 8-10.

4 Positive percent agreement (PPA) = 100xTP/(TP+FN) where the true positives (TP) are the number of positive variant calls at genomic coordinates where variants are

{14}------------------------------------------------

present according to the reference sequence and mutant allele called is concordant with reference sequence (column named "Variants called correctly") and the false negatives (FN) are the number of negative variant calls at genomic coordinates where variants are present according to the reference sequence (column named "Variants missed).

5 Negative percent agreement (NPA) = 100xTN/(FP+TN) where the false positives (FP) are the number of positive variant calls at genomic coordinates where variants are absent according to the reference sequence, or if mutant allele called is discordant with reference sequence (not in the table; no false positive variants calls were made in this study) and true negatives (TN) are the number of negative variant calls at genomic coordinates where variants are absent according to the reference standard (column named "non-variant bases called correctly").

Study 2: The sequencing results for the amplicon panel above were compared to a highly confident genotype established for NA12878 by the National Institutes of Standards and Technology (NIST) (v.2.152). Out of the 195 amplicons, 184 amplicons lied within highly confident reference calls in the NIST sequence and were compared. Non-variant base calls were compared to human genome reference sequence build 19.

Sample#Amplicons%Amplicon1CoverageVariantsexpectedVariantsCorrectlyCalledVariantsMissedNon-variantbases calledcorrectlyPPA2(%)NPA3(%)
NA128781841001716142306694.1100

1 % Amplicon coverage is number of bases in the amplicons sequenced with confidence

2 Positive percent agreement (PPA) = 100xTP/(TP+FN)

3 Negative percent agreement (NPA) = 100xTN/(FP+TN)

4 The missed variant is the one base pair deletion in amplicon 9 in the homopolymer run of 14 A's not called by the MiSeqDx that is present in the NIST sequence. Note that the NIST sequence does not include the one base pair insertion in the other homopolymer of A's that was present in the other reference database used above in study 1.

2 Zook, JM et al. Integrating sequencing datasets to form highly confident SNP and indel genotype calls for a whole human genome. arXiv:1307.4661 [q-bio.GN]

{15}------------------------------------------------

Study 3: An additional accuracy study was performed to assess the performance of small insertions and deletions within a representative assay, the Illumina MiSeqDx Cystic Fibrosis 139 Variant Assay, that included a subset of CFTR clinically significant genetic variations analyzed with the MiSeq Reporter (MSR) software v2.2.29.2 using the MiSeqDx Platform targeted DNA sequencing workflow. The queried insertions and deletions were detected where expected with high confidence. These samples were characterized by bidirectional Sanger sequencing as a reference method to establish the expected sequence.

AmpliconAnalyzed fragment size1Amplicon Genomic Content# calls/ sample that could be made# of bases called/ sample# of no calls# of correct calls/ sample# of incorrect calls% correct calls
11291 base insertion13013001300100.00
21543 base deletion15115101510100.00
31672 base deletion16516501650100.00
41341 base deletion13313301330100.00
51321 base deletion13113101310100.00
61291 base deletion12812801280100.00

The data provided by these three accuracy studies supports the claim that the MiSeqDx Instrument can accurately sequence:

  • GC content > 19% (all bases in 135 out of 135 sequenced amplicons with 19% GC 0 content called correctly)
  • o GC content < 72% (all bases in 135 out of 135 sequenced amplicons with 72% GC content called correctly)
  • 0 PolyA lengths ≤ 7 (PolyA repeat of 7 nucleotides was called correctly in 270 out of 270 sequenced amplicons containing PolyA =7)
  • PolyT lengths ≤ 8 (PolyT repeat of 8 nucleotides was called correctly in 270 out of 270 sequenced amplicons containing PolyT =8)
  • PolyG lengths ≤ 6 (PolyG repeat of 6 nucleotides was called correctly in 405 out of 405 sequenced amplicons containing PolyG =6)
  • PolyC lengths ≤ 7 (PolyC repeat of 7 nucleotides was called correctly in 135 out of 135 o sequenced amplicons containing PolyC =7) )
  • o Dinucleotide repeat lengths < 5x (all bases in 135 out of 135 sequenced amplicons with 5x dinucleotide repeat were called correctly)
  • Trinucleotide repeat lengths ≤ 4x (all bases in 810 out of 810 sequenced amplicons with 0 4x trinucleotide repeats were called correctly)
  • 1 base insertions and 3 or fewer base deletions

{16}------------------------------------------------

  • o 2 out of 3 1-base insertions tested were called correctly. Correct calls were made for two 1-base insertions in non-homopolymer regions in 82 amplicons. One 1-base insertion was not called in a homopolymer run of 14 A's on chromosome 2.in 135 amplicons.
  • 3 out of 4 1-base deletions called correctly. All correct calls were made in o non-homopolymer regions in 4 amplicons. One 1-base deletion was not called in a homopolymer run of 14 A's on chromosome 9 in 63 amplicons.
  • 2-base deletion called correctly in one sample о
  • o 3-base deletions called correctly in 21 samples

c. Precision/Reproducibility:

Two reproducibility studies were conducted.

Study 1: This reproducibility study used a representative assay designed to query a variety of genes covering 24,434 bases across 19 different chromosomes, and containing potentially clinically relevant exons. The study examined 13 samples over nine runs using three different MiSeqDx instruments and three different operators. The 13 samples are from two parents and 11 children that have been frequently sequenced by multiple laboratories and sequencing methodologies. The samples were run in duplicate, so each run generated results for 15 samples.

For the evaluation of lot-to-lot reproducibility, 94 samples and two non-template controls were tested across three lots. Each lot was split into two 48-sample runs to test all reagents and possible index primer combinations. All sequencing runs were completed by a single operator and on a single MiSeqDx instrument to remove any potential variance contributed from operator or instrument.

Correct calls were determined for single nucleotide variants (SNVs) by comparing the study data to well characterized reference information. No repeat testing was done for the reproducibility study. The following tables show the results of the study to evaluate reproducibility of the system.

AmpliconChr.Analyzed fragment size1Amplicon Genomic Content# of samples2 runMiSeqDx 1MiSeqDx 2MiSeqDx 3
total # of no calls3total # of incorrect calls4% correct calls5total # of no callstotal # of incorrect calls% correct callstotal # of no callstotal # of incorrect calls% correct calls
11132Poly C (5); 63% GC13500100.00236099.617396099.34
21128Poly T (5)13500100.0000100.0000100.00
32133-13500100.0000100.0000100.00
Amp-liconChr.Analyzedfragmentsize1AmpliconGenomicContent# ofsamplesrun2MiSeqDx 1total #of nocalls3total # ofincorrectcalls4%correctcalls5MiSeqDx 2total #of nocalls3total # ofincorrectcalls4%correctcalls5MiSeqDx 3total #of nocalls3total # ofincorrectcalls4%correctcalls5
42119-13500100.0000100.0000100.00
52127Poly T (5)13500100.0000100.0000100.00
62135Poly A (6)13500100.0000100.0000100.00
72122Poly T (5);Poly C (5)13500100.0000100.0000100.00
82110Poly T (5)13500100.0000100.0000100.00
92131Poly A (14)135027899.54027899.54027899.54
102117-13500100.0000100.0000100.00
112121-13500100.0000100.0000100.00
122114-13500100.0000100.0000100.00
132129Poly A (5)13500100.0000100.0000100.00
143131Poly A (5);Poly T (5)13500100.0000100.0000100.00
153130-13500100.0000100.0000100.00
163130-13500100.0000100.0000100.00
173117-13500100.0000100.0000100.00
183136Poly T (5)13500100.0000100.0000100.00
193131Poly T (5);SNV13500100.0000100.0000100.00
203123Poly A (5)13500100.0000100.0000100.00
213117Poly A (6);Poly T (5);Homologousregion on adifferentchromosome13500100.0000100.0000100.00
223119Homologousregion on adifferentchromosome13500100.0000100.0000100.00
233120-13500100.0000100.0000100.00
243129Poly T (5)13500100.0000100.0000100.00
254133Poly C (7);66% GC13500100.0000100.0000100.00
264135Poly C (5);69% GC13500100.0000100.0000100.00
274123SNV13500100.0000100.0000100.00
284134-13500100.0000100.0000100.00
294132-13500100.0000100.0000100.00
Ana-lyzedfragmentsize¹AmpliconGenomicContent# ofsamplesrun²MiSeqDx 1MiSeqDx 2MiSeqDx 3
Amp-liconChr.total #of nocalls³total # ofincorrectcalls⁴%correctcalls⁵total #of nocalls³total # ofincorrectcalls%correctcalls⁵total #of nocalls³total #ofincorrectcalls%correctcalls⁵
304121Poly A (5);SNV13500100.0000100.0000100.00
314125-13500100.0000100.0000100.00
324134Poly T (5)13500100.0000100.0000100.00
334118-13500100.0000100.0000100.00
344122Poly A (5)13500100.0000100.0000100.00
354131-13500100.0000100.0000100.00
364133-13500100.0000100.0000100.00
374128Poly T (6)13500100.0000100.0000100.00
384131-13500100.0000100.0000100.00
394129Poly A (5);Poly T (5);SNV13500100.0000100.0000100.00
404133Poly T (5);SNV13500100.0000100.0000100.00
414112SNV13500100.0000100.0000100.00
424133-13500100.0000100.0000100.00
434135-13500100.0000100.0000100.00
444122-13500100.0000100.0000100.00
454117-13500100.0000100.0000100.00
464124-13500100.0000100.0000100.00
474117Poly T (5)13500100.0000100.0000100.00
484128Poly A (7)13500100.0000100.0000100.00
494123Poly A (6)13500100.0000100.0000100.00
504133-13500100.0000100.0000100.00
514112-13500100.0000100.0000100.00
524129-13500100.0000100.0000100.00
534126-13500100.0000100.0000100.00
544132-13500100.0000100.0000100.00
555131-13500100.0000100.0000100.00
565119-13500100.0000100.0000100.00
575120Poly A (5)13500100.0000100.0000100.00
585119-13500100.0000100.0000100.00
595118-13500100.0000100.0000100.00
605112-13500100.0000100.0000100.00
615120-13500100.0000100.0000100.00
625120Poly A (5)13500100.0000100.0000100.00
635115CT(5)13500100.0000100.0000100.00
Amp-liconChr.Analyzedfragmentsize¹AmpliconGenomicContent# ofsamplesrun²MiSeqDx 1MiSeqDx 2MiSeqDx 3
total #of nocalls³total # ofincorrectcalls⁴%correctcalls⁵total #of nocalls³total # ofincorrectcalls%correctcalls⁵total #of nocallstotal #ofincorrectcalls%correctcalls
645112SNV13500100.0000100.0000100.00
655135Poly T (6)13500100.0000100.0000100.00
66513163% GC13500100.0000100.0000100.00
675121-13500100.0000100.0000100.00
685132Poly A (6);Poly T (8)13500100.0000100.0000100.00
697133-13500100.0000100.0000100.00
70712060% GC13500100.0000100.0000100.00
717135-13500100.0000100.0000100.00
727126Poly A (5);59% GC13500100.0000100.0000100.00
737134-13500100.0000100.0000100.00
747122Poly C (5);63% GC13500100.0000100.0000100.00
75712759% GC;SNV13500100.0000100.0000100.00
767123-13500100.0000100.0000100.00
777125-13500100.0000100.0000100.00
787133Poly A (5);Poly T (5)13500100.0000100.0000100.00
797116-13500100.0000100.0000100.00
807135-13500100.0000100.0000100.00
817118-13500100.0000100.0000100.00
82713667% GC13500100.0000100.0000100.00
83713158% GC13500100.0000100.0000100.00
847119Poly G (6);61 GC13500100.0000100.0000100.00
857122Poly T (5)13500100.0000100.0000100.00
867123Poly A (6)13500100.0000100.0000100.00
87812760% GC13500100.0000100.0000100.00
88812957% GC13500100.0000100.0000100.00
899130Poly T (5)13500100.0000100.0000100.00
909116-13500100.0000100.0000100.00
919119Homologousregion on adifferentchromosome13500100.0000100.0000100.00
929121-13500100.0000100.0000100.00
Amp-liconChr.Analyzedfragmentsize¹AmpliconGenomicContent# ofsamplesrun²total #of nocalls³total # ofincorrectcalls⁴%correctcalls⁵total #of nocallstotal # ofincorrectcalls%correctcallstotal #of nocallstotal #ofincorrectcalls%correctcalls
939117Homologousregion on adifferentchromosome13500100.0000100.0000100.00
94911413500100.0000100.0000100.00
959129Poly A (14)135045⁹99.22045⁹99.22045⁹99.22
969114Homologousregion on adifferentchromosome; SNV13500100.0000100.0000100.00
979122-13500100.0000100.0000100.00
989127Poly A (5);Poly C (5)13500100.0000100.0000100.00
999133-13500100.0000100.0000100.00
100913864% GC13500100.0000100.0000100.00
1019139-13500100.0000100.0000100.00
1029116-13500100.0000100.0000100.00
1039133Poly A (5);57% GC13500100.0000100.0000100.00
104913857 GC13500100.0000100.0000100.00
1059136Poly C (5);67% GC13500100.0000100.0000100.00
106911870% GC13500100.0000100.0000100.00
1071012862% GC13500100.0000100.0000100.00
1081012060% GC13500100.0000100.0000100.00
1091013958% GC;SNV13500100.0000100.0000100.00
1101011857% GC13500100.0000100.0000100.00
11110123Poly T (5)13500100.0000100.0000100.00
11210121-13500100.0000100.0000100.00
1131012926% GC13500100.0000100.0000100.00
11410122-13500100.0000100.0000100.00
11510124Poly T (5);Homologousregion on adifferentchromosome13500100.0000100.0000100.00
11610135CA (4)13500100.0000100.0000100.00

Results of the study by instrument:

{17}------------------------------------------------

{18}------------------------------------------------

{19}------------------------------------------------

{20}------------------------------------------------

{21}------------------------------------------------

Ana-MiSeqDx 1MiSeqDx 2MiSeqDx 3
Amp-lyzedAmplicon# oftotal # total # of0/0total #total # of%total #total #%
liconChr.frag-Genomicsamplesof noincorrectcorrectof noincorrectcorrectof noofcorrect
mentContentrun"calls³calls"calls®callscallscallscallsincorrectcalls
sizecalls
11710ા રેરPoly A (6);ા રેર00100.0000100.0000100.00
Homologous
region on adifferent
chromosome
11810119Poly C (5);ા રેડ00100.0000100.0000100.00
SNV
11910125ાં રેણ00100.0000100.0000100.00
12010131ા રેડ00100.0000100.0000100.00
12110117ા રેડ00100.0000100.0000100.00
12210116ાં રેણ00100.0000100.0000100.00
1231012958% GCા રેર00100.0000100.0000100.00
12411117Poly T (10)ા રેર00100.0000100.0000100.00
ા ટર્ડ11117Poly T (5)ા રેર00100.0000100.0000100.00
12611113Poly A (5)ા રેર00100.0000100.0000100.00
12711129ા રેર00100.0000100.0000100.00
12811121Poly T (5)ાં રેણ00100.0000100.0000100.00
12911123ાં રેણ00100.0000100.0000100.00
13011127Poly A (6)ા રેર00100.0000100.0000100.00
13111136Poly T (6)ાં રેણ00100.0000100.0000100.00
13211132Poly T (5)ાં રેણ00100.0000100.0000100.00
13311ા રિાં રેણ00100.0000100.0000100.00
134l I117Poly T (8):19% GCા રેર00100.0000100.0000100.00
ા રેર11134Poly A (5);ા રેર00100.0000100.0000100.00
Poly T (5)
13611131Poly A (5)ા રેર00100.0000100.0000100.00
13711133SNV: 26%ાં રેણ00100.0000100.0000100.00
GC
13811137Poly T (8);SNVા રેર00100.0000100.0000100.00
13911131Poly A (5)ા રેર00100.0000100.0000100.00
14012131ાં રેર00100.0000100.0000100.00
14112128ાં રેર00100.0000100.0000100.00
14212133Poly A (5)ા રેર00100.0000100.0000100.00
14312136ા રેર00100.0000100.0000100.00
14412124। ૩૨00100.0000100.0000100.00
1451212259% GCા રેર00100.0000100.0000100.00
Ana-lyzedfrag-mentsize¹AmpliconGenomicContent# ofsamplesrun²MiSeqDx 1MiSeqDx 2MiSeqDx 3
Amp-liconChr.total #of nocalls³total # ofincorrectcalls⁴%correctcalls⁵total #of nocalls³total # ofincorrectcalls%correctcalls⁵total #of nocallstotal #ofincorrectcalls%correctcalls
14613122-13500100.0000100.0000100.00
14713116Poly C (5)13500100.0000100.0000100.00
14813133-13500100.0000100.0000100.00
14913117SNV13500100.0000100.0000100.00
15013124Poly T (6)13500100.0000100.0000100.00
15113123Poly T (5);26% GC13500100.0000100.0000100.00
15213115Poly A (5)13500100.0000100.0000100.00
15313125-13500100.0000100.0000100.00
15413121-13500100.0000100.0000100.00
15513123-13500100.0000100.0000100.00
15613114-13500100.0000100.0000100.00
15713119-13500100.0000100.0000100.00
1581412258% GC13500100.0000100.0000100.00
15916122-13500100.0000100.0000100.00
16016121-13500100.0000100.0000100.00
16116123Poly C (5)13500100.0000100.0000100.00
16217119-13500100.0000100.0000100.00
1631711961% GC13500100.0000100.0000100.00
16417135-13500100.0000100.0000100.00
16517116Poly C (6);60% GC;SNV13500100.0000100.0000100.00
16617123-13500100.0000100.0000100.00
1671711662% GC13500100.0000100.0000100.00
16817118Poly C (5);65% GC13500100.0000100.0000100.00
16917129-13500100.0000100.0000100.00
17017131Poly G (6);67% GC;SNV13500100.0000100.0000100.00
1711712761% GC13500100.0000100.0000100.00
17217118Poly C (5)13500100.0000100.0000100.00
1731713861% GC13500100.0000100.0000100.00
1741713158% GC13500100.0000100.0000100.00
17518112-13500100.0000100.0000100.00
17618124-13500100.0000100.0000100.00
17718134Poly A (6)13500100.0000100.0000100.00
MiSeqDx 1MiSeqDx 2MiSeqDx 3
Amp-liconChr.Analyzedfragmentsize¹AmpliconGenomicContent# ofsamplesrun²total #of nocalls³total # ofincorrectcalls⁴%correctcalls⁵total #of nocalls³total # ofincorrectcalls%correctcalls⁵total #of nocalls³total #ofincorrectcalls%correctcalls⁵
17818129-13500100.0000100.0000100.00
17918133-13500100.0000100.0000100.00
18018118-13500100.0000100.0000100.00
1811811460% GC13500100.0000100.0000100.00
18218118-13500100.0000100.0000100.00
18319122Poly G (6);66% GC13500100.0000100.0000100.00
1841913964% GC13500100.0000100.0000100.00
1851913167% GC13500100.0000100.0000100.00
1861914159% GC;Homologousregion on adifferentchromosome13500100.0000100.0000100.00
18719121Poly C (5);72% GC;Homologousregion on adifferentchromosome13500100.0000100.0000100.00
1881913858% GC13500100.0000100.0000100.00
1891912364% GC13500100.0000100.0000100.00
19019138-13500100.0000100.0000100.00
19120117Poly T (5)13500100.0000100.0000100.00
19222136Poly A (7)13500100.0000100.0000100.00
19322122Poly A (5);Poly C (5)13500100.0000100.0000100.00
1942212262% GC;SNV13500100.0000100.0000100.00
1952211966% GC13500100.0000100.0000100.00

{22}------------------------------------------------

{23}------------------------------------------------

1 Analyzed fragment size means the size of the sequenced genomic region in bases, not including target-specific primers.

2 Number of samples is calculated from 9 runs of 15 samples (11 samples run once and 2 samples run twice)

3 Total number of no calls is the combined number of no calls obtained for all 135 runs at that amplicon

4 Total number of incorrect calls is the combined number of incorrect calls obtained

{24}------------------------------------------------

for all 135 runs at that amplicon

5 % correct calls equals the correct call rate for all of the bases in the amplicon, where the correct call for the SNV or indel is based on the well characterized reference database and the correct call for the bases in the remainder of the amplicon sequence is based on comparison to . This column may have more than one expected result for a given amplicon if some samples are expected to have an indel and some are not, e.g., amplicon 9.

6 Amplicon 1 had a number of bases whose genotype could not be called: 12 bases in 1/9 runs in NA12881; 1 base in 2/9 runs and 3 bases in 1/9 runs in NA12886; 20 bases in 1/9 runs and 26 bases in 1/9 runs in NA12888. This is due to low coverage at no call bases in those runs, where the average sequencing depth was 33.2, with a minimum of 21 and maximum of 52.

7 When no-calls are not included in the calculation, the correct call rate is 100%.

8 Amplicon 9 has a homopolymer run of 14 A's according to the human genome reference sequence build 19. However, the well characterized reference information for 7 out of 13 samples have 13 A's in this homopolymer run. In these 7 samples, this one base pair deletion is called a false negative, and is called as false negative reproducibly in all nine runs..

9 Amplicon 95 has a homopolymer run of 14 A's according to human genome reference sequence build 19. . However, the well characterized reference information sequences for 13 out of 13 samples have 15 A's in this homopolymer run. In these 13 samples, this one base pair insertion is 100% reproducibly not called (i.e., it is false negative).

Study 1 results below present results each sample compounded from all nine runs into one column. The results displayed are solely for the single nucleotide variants and insertions/deletions results versus the reference database sequence. This analysis demonstrated that the results for the variants were reproducible across nine runs for these samples.

DNA #DNA Sample ID
1NA12877
2NA12878
3NA12879
4NA12880
5NA12881
6NA12882
7NA12883

Sample panel indicating sample number for the study and associated sample name/ID:

{25}------------------------------------------------

DNA #DNA Sample ID
8NA12884
9NA12885
10NA12886
11NA12887
12NA12888
13NA12893

Reproducibility results for SNVs and Indels per sample:

DNA## RunspersampleSingle Nucleotide Variants (SNVs)Insertions\Deletions (Indels)
# ofSNVs# CalledCorrectly# of FalsePositives1# of FalseNegatives2# ofIndels# CalledCorrectly# of FalsePositives1# of FalseNegatives2
13181616003102
23181717002002
391818003101
491717003102
591919003102
691515001001
792222002101
891919002101
991717002002
1091919003102
1191818001001
1292222002101
1391717003102

4 False Positive = Variant called by MiSeqDx sequencing run but not in reference database

2 False Negative = Variant in reference database but not called in MiSeqDx sequencing run.

3 Samples NA12877 and NA12878 were run in duplicate. Replicate samples generated identical results

Study2: A reproducibility study performed with a representative assay, the Illumina MiSeqDx Cystic Fibrosis 139 Variant Assay, included a subset of CFTR clinically significant genetic variations analyzed with the MiSeq Reporter (MSR) software v2.2.29.2 using the MiSeqDx Platform targeted DNA sequencing workflow. The blinded study used 3 trial sites and 2 operators at each site. Two well-characterized panels of 46 samples each were tested by each of the operators at each site for a total

{26}------------------------------------------------

of 810 calls per site. The panels contained a mix of genomic DNA from cell lines with known variants in the CFTR gene, as well as leukocyte-depleted blood spiked with cell lines with known variants in the CFTR gene. The blood samples were provided to allow incorporation of the extraction steps used to prepare gDNA that serves as the primary input for the assay workflow. The sample pass rate, defined as the number of samples passing QC metrics on the first attempt, was 99.9%. All test results are based on initial testing. No repeat testing was done for the reproducibility study.

Sample GenotypeTotal calls per sitePositive Agreeing Calls (Variants) Site 1Positive Agreeing Calls (Variants) Site 2Positive Agreeing Calls (Variants) Site 3Negative Agreeing Calls (Wild type) Site 1Negative Agreeing Calls (Wild type) Site 2Negative Agreeing Calls (Wild type) Site 3# Miscalls# No CallsPositive Agreement (%)Negative Agreement (%)Overall Agreement (%)
S549N (HET)81066680480480400100100100
1812-1 G->A (HET)81066680480480400100100100
Q493X/F508del (HET)81012121279879879800100100100
F508del/2184delA (HET)81012121279779879801100100100
Y122X/R1158X (HET)2810121112798664798013597.2294.4094.44
F508del/2183AA>G81012121279879879800100100100
R75X (HET)81066680480480400100100100
I507del/F508del (HET)81012121279879879800100100100
F508del/W1282X (HET)38101211127987977982097.2299.9699.9
F508del/3272-26A>G (HET)38101211127987977982097.2299.9699.9
F508del/3849+10kbC>T (HET)81012121279879879800100100100
621+1G>T/3120+1G>A (HET)81012121279879879800100100100
E60X/F508del (HET)81012121279879879800100100100
M1101K (HET)81066680480480400100100100
M1101K (HOM)81066680480480400100100100
F508del (HOM)482866682282282200100100100
F508del/3659delC (HET)81012121279879879800100100100
TotalcallsPositive AgreeingCalls (Variants)Negative AgreeingCalls (Wild type)#PositiveNegativeOverall
Sample GenotypepersiteSitelSite2Site3SitelSite2Site3MiscallS# NoCallst (%)Agreemen Agreement Agreement(%)(%)
R117H/F508del (HET) ³81618181879879879800100100100
621+1G>T/711+1G>T(HET)81012121279879879800100100100
G85E/621+1G>T (HET)81012121279879879800100100100
A455E/F508del (HET)81012121279879879800100100100
F508del/R560T (HET)81012121279879879800100100100
F508del/Y1092X (C>A)(HET)81012121279879879800100100100
N1303K (HET)81066680480480400100100100
G542X (HOM)81066680480480400100100100
G542X (HET)81066680480480400100100100
G551D/R553X (HET)81012121279879879800100100100
3849+10kbC>T (HOM)81066680480480400100100100
WT81000081081081000N/A100100
F508del (HET)81066680480480400100100100
1717-1G>A (HET)81066680480480400100100100
R1162X (HET)81066680480480400100100100
R347P/G551D (HET)81012121279879879800100100100
R334W (HET)81066680480480400100100100
WT81000081081081000N/A1001 00
G85E (HET)81066680480480400100100100
I336K (HET)81066680480480400100100100
WT81000081081081000N/A100100
F508de1/3849+10kbC>T(HET)81012121279879879800100100100
621+1G>T/3120+1G>A(HET)81012121279879879800100100100
F508del/3659de1C(HET)81012121279879879800100100100
R117H/F508del (HET)S81618181879879879800100100100
G85E/621+1G>T (HET)81012121279879879800100100100
A455E/F508del (HET)81012121279879879800100100100
TotalcallsPositive AgreeingCalls (Variants)Negative AgreeingCalls (Wild type)#PositiveNegativeOverall
Sample GenotypeperSite1Site2Site3Site1Site2Site3Miscall# NoCallsAgreement(%)Agreement(%)Agreement(%)
sites
N1303K (HET)81066680480480400100100100
G551D/R553X (HET)81012121279879879800100100100
2789+5G>A (HOM)81066680480480400100100100
F508del/1898+1G>A(HET)81012121279879879800100100100
WT81000081081081000N/A100100
F508del/2143delT(HET)81012121279879879800100100100
3876delA (HET)81066680480480400100100100
3905insT (HET)81066680480480400100100100
394delTT (HET)81066680480480400100100100
F508del (HET)81066680480480400100100100
WT81000081081081000N/A100100
WT81000081081081000N/A100100
F508del (HET)81066680480480400100100100
WT81000081081081000N/A100100
L206W (HET)81066680480480400100100100
WT81000081081081000N/A100100
G330X (HET)81066680480480400100100100
WT81000081081081000N/A100100
R347H (HET)81066680480480400100100100
1078delT (HET)81066680480480400100100100
G178R/F508del (HET)81012121279879879800100100100
S549R (c.1647T>G)(HET)81066680480480400100100100
S549N (HET)81066680480480400100100100
W846X (HET)81066680480480400100100100
WT81000081081081000N/A100100
E92X/F508del (HET)81012121279879879800100100100
621+1G>T/1154insTC(HET)8101212127987987970110099.9699.96
G542X (HET)81066680480480400100100100
Sample GenotypeTotalcallspersiteSite1Positive AgreeingCalls (Variants)Site2Site3SitelNegative AgreeingCalls (Wild type)Site2Site3#Miscallഗ# NoCallsPositivet (%)Negative(%)OverallAgreemen Agreement Agreement(%)
F508del (HET)81066680480480400100100100
F508del (HET)-8106568046708040ા ૩૨94.4494.4494.44
F508del (HET)81066680480480400100100100
621+1G>T/A455E(HET)81012121279879879800100100100
1812-1 G->A (HET)81066680480480400100100100
WT81000081081081000N/A100100
F508del/R553X (HET)81012121279879879800100100100
F508del/G551D (HET)81012121279879879800100100100
R347P/F508del (HET)810121212798798798001001 00100
R117H/F508del (HET)81618181879879879800100100100
I507del (HET)81066680480480400100100100
2789+5G>A (HOM)81066680480480400100100100
F508del/1898+1G>A(HET)81012121279879879800100100100
WT81000081081081000N/A100100
F508del/2143delT(HET)81012121279879879800100100100
3905insT (HET)81066680480480400100100100
394delTT (HET)81066680480480400100100100
F508del (HET)81066680480480400100100100

{27}------------------------------------------------

{28}------------------------------------------------

{29}------------------------------------------------

¹ Variant N1303K was not called in this sample

2 One replicate of samples 5 and 75 had a 0% call rate. Further investigation indicates that samples may not have been added to the sample plate prior to library preparation

3 Evidence indicates that samples 9 and 10 were likely switched by the operator prior to library preparation.

4 I506V, I507V, F508C not present in sample

5 Sample also has the (TG)10 (T)9/(TG) 12(T)5 variant

6 Variant M1V was not called in these samples

c. Linearity:

{30}------------------------------------------------

Not applicable.

d. Carryover:

A study was performed to evaluate the potential for inter-run and intra-run sample carryover. Intra-run sample carryover tested the system in the most challenging scenario for sample carryover within a single sequencing run. One 48-sample library composed of two samples with unique variants was setup in a checkerboard matrix pattern at alternating high (500 ng) and low (100 ng) concentrations, along with 4 NTC's (no template control samples).

Inter-run sample carryover tested the system for sample carryover between successive sequencing runs. Two libraries were prepared; each library was composed of 47 replicates of a single genomic DNA sample and one NTC. Each library used a different sample from the other.

For both inter- and intra-run, sample carryover was determined by measuring the error rate at the position of variant calls for all samples used in the study. Acceptance criteria for this study were reviewed and deemed acceptable. This study met the sponsor's acceptance criteria and demonstrated that there is minimal to no carryover on the MiSeqDx.

e. Interfering Substances:

To assess the impact of interfering substances on the MiSeqDx Platform, a representative assay (Assay 2) was evaluated in the presence and absence of potential interferents. Eight whole blood samples representing eight unique genotypes were utilized in the study. Four endogenous interfering substances (bilirubin, cholesterol, hemoglobin, and triglycerides) were tested by spiking them into the blood specimens prior to DNA extraction. To assess interference resulting from blood collection (short draw), potassium EDTA was spiked into blood samples at two concentrations. The concentration limits for each substance is shown in the following table. Additionally, to assess interference resulting from sample preparation, 15% wash buffer was added to 8 purified genomic DNA with 100% correct calls.

{31}------------------------------------------------

TestSubstanceTotal Numberof ReplicatesConcentrationTested in Blood(Upper Limit)ConcentrationTested in Blood(Lower Limit)CallRate
Bilirubin48684 µmol/L137 µmol/L100%
Cholesterol4813 mmol/L2.6 mmol/L100%
Hemoglobin482 g/L0.4 g/L100%
Triglyceride4837 mmol/L7.4 mmol/L100%
EDTA487.0 mg/mL2.8 mg/mL100%

2. Other Supportive Instrument Performance Data Not Covered Above:

DNA extraction study: Three different extraction methods, magnetic bead extraction, alcohol precipitation and silica filter column isolation were evaluated using K2EDTA anticoagulated whole blood. Fourteen unique blood samples were used in the study representing a range of genotypes from one representative gene. The three DNA extraction methods were tested independently by 2 different operators who each performed 3 runs per extraction method. Each extraction was performed by each operator on different days. The DNA concentration and A260/A280 ratio of the extracted gDNA samples was determined using spectrophotometry. The total sample size for each extraction method in this study was 168 (14 samples x 2 operators/extraction method x 3 runs/operator x 2 replicates/extracted gDNA sample).

Extraction MethodNumber ofsamples testedCall RateAccuracySample FirstPass Rate*
Alcohol Precipitation(Qiagen Gentra PureGene)168100%100%100%
Silica Filter ColumnIsolation (Qiagen BloodMini)168100%100%100%
Magnetic Bead extraction(Biomerieux Easy Mag)168100%100%100%

{32}------------------------------------------------

DNA input study: The DNA input range for the MiSeqDx Platform was evaluated by performing a serial dilution study using 14 representative DNA samples containing 16 unique single gene variants. Each sample was tested in duplicate at 9 DNA input levels ranging from 1250 ng to 1 ng (1250 ng, 500 ng, 250 ng, 100 ng, 50 ng, 10 ng, 5 ng, and 1 ng). For determination of accuracy, sample genotypes were compared to bidirectional Sanger sequencing data and the deletions were compared to PCR assay. 1250 ng and 25 ng were identified as the upper and lower bound for DNA input respectively as they had ≥95% sample first pass rate with no incorrect calls (100% accuracy and call rate).

DNA inputs of 1250 ng, 250 ng, and 100 ng were further tested with 4 representative DNA samples and 20 replicates per DNA input level for each sample (n=420=80 samples), while the lower bound of 25 ng was tested with 14 samples, 20 replicates for each sample (n=1420=280 samples). The accuracy and sample first pass rate was 100% at all DNA input levels. There were 2 no calls overall observed at the 25 ng DNA input level, with sample call rates of 99.26%.

Thermal cycler study: Three different commercially available thermal cyclers were evaluated using the representative assay, the Illumina MiSeaDx Cystic Fibrosis 139 Variant Assay. Thermal cycles are used in the library preparation. Three unique sample sets were processed through all three thermal cyclers across 3 days. This enabled performance assessment across different thermal cyclers on different days. Each sample set was processed in triplicate each day (i.e. one replicate per thermal cycler). Acceptance criteria for this study were reviewed and deemed acceptable. This study met the sponsor's acceptance criteria and demonstrated that any commercially available thermal cycler would be adequate for library preparation for use with the MiSeqDx.

Sample indexing study: Sample index primers are used in the kit to assign a unique barcode to each sample DNA, allowing the ability to pool multiple samples together into a single sequencing run.

A total of 96 samples indexes were tested with Assay 2 using 8 unique DNA samples to verify the ability of the assay to consistently make a genotyping call for a given sample across different indexing primer combinations. Each sample was tested with 12 different indexing primer combinations. Sample results were compared against bidirectional Sanger sequencing data for all positions/variants. Reproducibility and accuracy were 100% for all sample/index primer combinations.

Specimen Storage: To verify the storage conditions and handling of blood samples for use with the MiSeqDx test system, six K2EDTA anti-coagulated blood samples were divided to six aliquots, one aliquot of each blood sample were stored under 6 different conditions: 2°C to 8°C for 1 day; -15°C to -25°C for 1 day; 2°C to 8°C for 30 days; -15°C to -25°C for 30 days: room temperature (20-25°C) for 7 days: and controlled room temperature (30°C) for 7 days. Genomic DNA was isolated from each aliquot using a commonly used commercial DNA extraction kit. All extractions were performed by a single operator. The extracted gDNA samples were stored at -15°C to -25°C until the libraries were prepared and sequenced.

{33}------------------------------------------------

The impact of repeated freeze-thaws on gDNA samples were tested by subjecting 15 DNA samples to 6 freeze thaw cycles.

Library preparations for both the samples from both the specimen storage and gDNA freeze-thaw studies were performed at the same time point. The samples from a single library preparation were pooled into one run of 48 samples and a second run of 32 samples prior to sequencing. Impact on call rate, reproducibility, and sample first pass rate were determined for each sample as compared to a respective control sample. No miscalls or no calls were observed for any of the specimens and demonstrated that the blood and gDNA storage conditions tested did not affect assay results.

K. Proposed Labeling:

Labeling satisfies the requirements of 21 CFR 809.10, 21 CFR 801.109, including an appropriate prescription statement as required by 21 CFR 801.109(b), and the special controls for this type of device.

L. Other Supportive Instrument Characteristics Data Not Covered In The "Performance Characteristics" Section above:

None.

Identified Potential RiskRequired Mitigation Measure
Inaccurate test results due to unavailabilityof necessary components of the instrumentsystemThe labeling for the instrument system mustreference pre-analytical and analytical reagents tobe used with the instrument system and include orreference legally marketed analytical softwarethat includes sequence alignment and variantcalling functions, to be used with the instrumentsystem.
Inaccurate results due to unknownperformance of the instrument systemThe labeling for the instrument system mustinclude a description of the followinginformation:i) The specimen type(s) validated asan appropriate source of nucleicacid for this instrument.ii) The type(s) of nucleic acids (e.g.,germline DNA, tumor DNA)validated with this instrument.iii) The type(s) of sequence variations(e.g. single nucleotide variants,insertions, deletions) validated withthis instrument.
Identified Potential RiskRequired Mitigation Measure
iv)The type(s) of sequencing (e.g.,targeted sequencing) validated withthis instrument.
v)The appropriate read depth for thesensitivity claimed and validationinformation supporting thoseclaims.
vi)The nucleic acid extractionmethod(s) validated for use withthe instrument system.
vii)Limitations must specify the typesof sequence variations that theinstrument cannot detect with theclaimed accuracy and precision(e.g., insertions or deletions largerthan a certain size, translocations).
viii)Performance characteristics of theinstrument system must include:
A)Reproducibility data generatedusing multiple instruments andmultiple operators, and atmultiple sites. Samples testedmust include all claimedspecimen types, nucleic acidtypes, sequence variationtypes, and types ofsequencing. Variants queriedshall be located in varyingsequence context (e.g.,different chromosomes, GC-rich regions). Device resultsshall be compared to referencesequence data with highconfidence.
B)Accuracy data for all claimedspecimen types and nucleicacid types generated by testinga panel of well-characterizedsamples to query all claimedsequence variation types,types of sequencing, andsequences located in varying
Identified Potential RiskRequired Mitigation Measure
sequence context (e.g.,different chromosomes, GC-rich regions). The well-characterized sample panelshall include samples from atleast two sources that havehighly confident sequencebased on well-validatedsequencing methods. At leastone reference source shallhave sequence generatedindependently of themanufacturer with respect totechnology and analysis.Percent agreement and percentdisagreement with thereference sequences must bedescribed for all regionsqueried by the instrument.
C) If applicable, data describingendogenous or exogenoussubstances that may interferewith the instrument system.
D) If applicable, datademonstrating the ability ofthe system to consistentlygenerate an accurate result fora given sample acrossdifferent indexing primercombinations.
ix) The upper and lower limit of inputnucleic acid that will achieve theclaimed accuracy andreproducibility. Data supportingsuch claims must also besummarized.

M. Identified Potential Risks and Required Mitigation Measures:

{34}------------------------------------------------

{35}------------------------------------------------

N. Benefit/Risk Analysis:

Summary

{36}------------------------------------------------

Summary ofthe Benefit(s)This is a tool for clinical laboratories that can provide accurate and reproduciblehigh throughput genomic sequencing of genomic regions of interest at greatersequencing depth than current sequencing technology. No other instruments are available for high throughput genomic sequence analysis.There is an unmet medical and public health need for a well-validated IVD labelledhigh throughput genomic sequence analyzer.
Summary ofthe Risk(s)Patients are subject to blood specimen collection, which is a standard procedure inclinical care and carries minimal risk. Risk is related to inaccurate test results as follows:False positive: The risks to the individual of a false positive result could includeunnecessary testing or treatment related to an inaccurate test result. Often, the resultfrom this test would be used with results from other diagnostic tests and clinicalsigns and symptoms to identify the genetic cause or contribution for a patient'sdisease or condition.False negative: The risks to the individual of a false negative result due to aninaccurate test result could delay further evaluation and appropriate therapy whichwill vary depending on the disease or condition.Public Health Risk from Incorrect Test Results:The consequences to public health for both false positive and false negative resultsare similar.
Summary ofOtherFactorsNot applicable.
ConclusionsDo the probable benefits outweigh the probable risks?

Given robust analytical performance characteristics and risk mitigation (i.e. extensive performance data provided in the labeling), the probable benefits to both the individual and public health outweigh the probable risks of this device.

0. Conclusion:

The information provided in this de novo submission is sufficient to classify this device into class II under regulation 21 CFR 862.2265. FDA believes that special controls, along with the applicable general controls, provide reasonable assurance of the safety and

{37}------------------------------------------------

effectiveness of the device type. This device, and similar devices, is classified under the following:

Product Code:PFF
Device Type:High throughput genomic sequence analyzer for clinical use
Class:II (special controls)
Regulation:21 CFR 862.2265

(a)Identification. A high throughput genetic sequence analyzer for clinical use is an analytical instrument system intended to generate, measure and sort signals in order to analyze nucleic acid sequences in a clinical sample. The device may include a signal reader unit: reagent handling, dedicated instrument control. and other hardware components: raw data storage mechanisms; data acquisition software; and software to process detected signals.

(b)Classification. Class II (special controls). A high throughput genetic sequence analyzer for clinical use must comply with the following special controls:

    1. The labeling for the instrument system must reference legally marketed pre-analytical and analytical reagents to be used with the instrument system and include or reference legally marketed analytical software that includes sequence alignment and variant calling functions, to be used with the instrument system.
    1. The labeling for the instrument system must Include a description of the following information:
    • i) The specimen type(s) validated as an appropriate source of nucleic acid for this instrument.
    • The type(s) of nucleic acids (e.g., germline DNA, tumor DNA) ii) validated with this instrument.
    • iii) The type(s) of sequence variations (e.g. single nucleotide variants, insertions, deletions) validated with this instrument.
    • iv) The type(s) of sequencing (e.g., targeted sequencing) validated with this instrument.
    • The appropriate read depth for the sensitivity claimed and v) validation information supporting those claims.
    • The nucleic acid extraction method(s) validated for use with vi) the instrument system.

{38}------------------------------------------------

  • vii) Limitations must specify the types of sequence variations that the instrument cannot detect with the claimed accuracy and precision (e.g., insertions or deletions larger than a certain size, translocations).
  • viii) Performance characteristics of the instrument system must include:
    • A) Reproducibility data generated using multiple instruments and multiple operators, and at multiple sites. Samples tested must include all claimed specimen types, nucleic acid types, sequence variation types, and types of sequencing. Variants queried shall be located in varying sequence context (e.g., different chromosomes, GC-rich regions). Device results shall be compared to reference sequence data with high confidence.
    • B) Accuracy data for all claimed specimen types and nucleic acid types generated by testing a panel of wellcharacterized samples to query all claimed sequence variation types, types of sequencing, and sequences located in varying sequence context (e.g., different chromosomes, GC-rich regions). The well-characterized sample panel shall include samples from at least two sources that have highly confident sequence based on well-validated sequencing methods. At least one reference source shall have sequence generated independently of the manufacturer with respect to technology and analysis. Percent agreement and percent disagreement with the reference sequences must be described for all regions queried by the instrument.
    • C) If applicable, data describing endogenous or exogenous substances that may interfere with the instrument system.
    • D) If applicable, data demonstrating the ability of the system to consistently generate an accurate result for a given sample across different indexing primer combinations.
    • ix) The upper and lower limit of input nucleic acid that will achieve the claimed accuracy and reproducibility. Data supporting such claims must also be summarized.

§ 862.2265 High throughput genomic sequence analyzer for clinical use.

(a)
Identification. A high throughput genomic sequence analyzer for clinical use is an analytical instrument system intended to generate, measure and sort signals in order to analyze nucleic acid sequences in a clinical sample. The device may include a signal reader unit; reagent handling, dedicated instrument control, and other hardware components; raw data storage mechanisms; data acquisition software; and software to process detected signals.(b)
Classification. Class II (special controls). The device is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to the limitations in § 862.9. The special controls for this device are:(1) The labeling for the instrument system must reference legally marketed pre-analytical and analytical reagents to be used with the instrument system and include or reference legally marketed analytical software that includes sequence alignment and variant calling functions, to be used with the instrument system.
(2) The labeling for the instrument system must include a description of the following information:
(i) The specimen type(s) validated as an appropriate source of nucleic acid for this instrument.
(ii) The type(s) of nucleic acids (
e.g., germline DNA, tumor DNA) validated with this instrument.(iii) The type(s) of sequence variations (
e.g. single nucleotide variants, insertions, deletions) validated with this instrument.(iv) The type(s) of sequencing (
e.g., targeted sequencing) validated with this instrument.(v) The appropriate read depth for the sensitivity claimed and validation information supporting those claims.
(vi) The nucleic acid extraction method(s) validated for use with the instrument system.
(vii) Limitations must specify the types of sequence variations that the instrument cannot detect with the claimed accuracy and precision (
e.g., insertions or deletions larger than a certain size, translocations).(viii) Performance characteristics of the instrument system must include:
(A) Reproducibility data generated using multiple instruments and multiple operators, and at multiple sites. Samples tested must include all claimed specimen types, nucleic acid types, sequence variation types, and types of sequencing. Variants queried shall be located in varying sequence context (
e.g., different chromosomes, GC-rich regions). Device results shall be compared to reference sequence data with high confidence.(B) Accuracy data for all claimed specimen types and nucleic acid types generated by testing a panel of well characterized samples to query all claimed sequence variation types, types of sequencing, and sequences located in varying sequence context (
e.g., different chromosomes, GC-rich regions). The well-characterized sample panel shall include samples from at least two sources that have highly confident sequence based on well-validated sequencing methods. At least one reference source shall have sequence generated independently of the manufacturer with respect to technology and analysis. Percent agreement and percent disagreement with the reference sequences must be described for all regions queried by the instrument.(C) If applicable, data describing endogenous or exogenous substances that may interfere with the instrument system.
(D) If applicable, data demonstrating the ability of the system to consistently generate an accurate result for a given sample across different indexing primer combinations.
(ix) The upper and lower limit of input nucleic acid that will achieve the claimed accuracy and reproducibility. Data supporting such claims must also be summarized.