Introducing NGS and Holotype HLA EAP User Experiences

Transcription

1

2 Agenda Introducing NGS and Holotype HLA EAP User Experiences David Lowe University of Liverpool Julio Delgado ARUP and University of Utah Alex Sheh Red Cross, Dedham Dimitri Monos CHOP Tim Hague Omixon EAP Results Whole Gene Consensus Q&A

3 Next-Generation Sequencing Recent technological advances, cumulatively called NGS, provide the first opportunity to fully characterize and fully phase the HLA genes Prof. Dimitri Monos, CHOP The Technique Long Range PCR Amplification Clonal sequencing of Amplicons Assembly of whole gene consensus sequences The Goal Automated, high-throughput analysis Unambiguous results No reflexive testing

4 Overview of NGS Sample Prep gdna extraction Long Range PCR Library Prep Size Selection Sequencing Paired-End 250bp Deep and even coverage High throughput Data Analysis Automated analysis Easy interpretation Whole gene consensus Unambiguous genotyping No reflexive testing

5 Introduction to Holotype HLA Holotype HLA is a combination Assay and Software product for the comprehensive gene amplification of multiple HLA loci, and sequencing on the Illumina MiSeq. NGS Assay + NGS Software

6 Omixon Holotype HLA 7 loci (Assay) 7 loci HLA-A, B, C, DRB1, DPB1, DQA1 and DQB1 X2 24/7 24 Indexes X2 96/7 96 Indexes Fully pooled (per sample indexing) 2 genotyping algorithms

7 Amplification Strategy HLA-A, B, C, DQA1 and DPA1 (Example is HLA-B) Full gene characterization 5 UTR UTR HLA-DQB1 5 UTR UTR Key region characterization HLA-DRB1, DRB3, DRB4 and DRB5 (example is DRB1) 5 UTR UTR HLA-DPB1 5 UTR UTR

8 Steps in Holotype HLA About 4 hours hands-on time for 24 samples (About 9 hours hands-on time for 96 samples)

9 Omixon HLA Twin (Software) Preconfigured Holotype HLA-specific settings Automated genotyping after MiSeq run Two algorithms for determining HLA genotypes Consensus Genotyping (Assembly) Statistical Genotyping (Alignment to IMGT/HLA) Traffic light system for data interpretation and workflow management

10 The Traffic Light System

11 Genotyping Results

12 Deep and Even Coverage

13 Reasons to choose Omixon Believe the Data routine clinical samples at CHOP 253 samples for ASHI validation study 2500+samples in worldwide EAP Technician friendly protocol All loci pooled together Only 1 bead cleanup 4 hours hands on time and 3.5 day TAT Automation Scalability samples per run 1, 5, 7 or 11 loci Entire team focused on NGS for HLA

14 Omixon Manufacture & Supply US-based contracted manufacturing ISO and GMP Operating Statistics 83,000 active customers 2,800 orders shipped per day 20,000 SKUs under management 100+ OEM product lines 4,500+ finished goods produced per month Worldwide distribution

15 Omixon Support and Services Presales Comprehensive NGS Cost Analysis Protocol viewing at an Omixon Reference Lab Hands on workshop at an Omixon Reference Lab Site Visit Consult/Presentation Site Preparation Full equipment, reagent and consumable check Evaluation and Training 1 week onsite training Data analysis walkthrough Ongoing remote and onsite support Validation 200 samples and 50 blind Complex case dataset Clinical Production Parallel testing

16 Early Access Program Worldwide Early Access Program (EAP) Launched at ASHI 2014 in Denver, CO Ran until end of June 2015 Early results presented at EFI 2015 in Geneva Full results announced today

17 Early Access Program Goals: Reproducibility Feedback and fine tuning Scope: 24 participants 13 countries samples

18 EAP Real Life 96-sample uncovered plate dropped on the floor just before sequencing = 1 Lab tech turned off thermal cycler on way out of the lab = 1 No LR Enzyme, protocol performed with honey & water = 1 Nano flow cell used by mistake = 1 Lab director nominated themselves for the first ever handson run through of the protocol = 1 Omixon's CEO visits lab during training = 2 Disaster follows = 2

19 Disasters solved by automation

20 EAP Guest Presenters David Lowe Liverpool University Hospital, UK

21 LIVERPOOL UNIVERSITY HOSPITAL OMIXON EARLY ACCESS USER REPORT

22 Where we re up to No previous sequencing experience in Liverpool To meet demand of expanding HSCT program A number of commercial partners considered

23 Really is early for us! 16 validation samples so far sequenced for HLA- A,B,C,DR,and DQ (2 runs of 8) Some problems along the way!! Helped to smooth out early issues- second run much smoother!

24 Data from Last Friday Thanks Tim for the Credits!!!

25 EAP Guest Presenters Julio Delgado ARUP

26 Evaluation of Omixon Holotype x2 24/7 kit for HLA typing: The ARUP Experience Julio Delgado, MD, MS Section Chief, Immunology Division, ARUP Laboratories, Salt Lake City Director, Histocompatibility and Immunogenetics Laboratory University of Utah Health Care, Salt Lake City Associate Professor, Department of Pathology, University of Utah School of Medicine, Salt Lake City

27 Enrolled in EAP - January 2015 On-site training by Omixon - April 2015 Evaluated 70 samples using Holotype X2-24/7 kit over 3 separate runs - May-June 2015 Data analysis - Jun-July 2015 Work performed by Tracie Profaizer in ARUP

28 Typing results of 70 individual samples (140 alleles/loci) A n=140 B n=140 C n=140 DPB1 n=140 DQA1 n=140 DQB1 n=140 DRB1 n=140 Total n=980 Unambiguous Typing (2 field) Failed (QC metric not met) 94.3% 94.3% 94.3% 90.0% 95.7% 97.1% 74.3% 91.4% 5.7% 5.7% 1.4% 5.7% 0.0% 1.4% 5.7% 3.7% Typing Failure 0% 0% 4.3% 4.3% 4.3% 1.4% 20.0% ** 4.9% **Due to weak amplification of DRB1 loci during first run

29 Summary Some typing failures in 1 st run due to inexperience, evaporation of some PCR reactions, and failure to appropriately use Promega Quantifluor results. Vast improvement with 2 nd and 3 rd runs. Recently, reanalyzed all data with latest software version, Twin 1.1, and improvement was seen across all samples. 100% concordance with Sanger reported alleles.

30 Highlights of Holotype x2 24/7 kit based on our experience The technical aspects of the kit are very easy to use. Index by sample. Pooling of loci early on in process reduces number of individual reactions to handle (and possibly mix-up). Pooling of all samples prior to final bead clean-up step, size selection, and quantification simplifies process even further. Excellent access to Omixon for guidance and troubleshooting Continued improvement of software for data analysis Product is the result of great coordination between R&D development by Dimitri s laboratory and Omixon s expertise in bioinformatics solutions

31 Future directions Need to continue to reduce number of repeats due to technical failures (evaporation, etc.). Continue to learn how to troubleshoot metric fails. Some metric fails are acceptable after review. Continue to familiarize ourselves with new version of Twin software. Comparison with other systems

32 EAP Guest Presenters Alex Sheh American Red Cross, Dedham

33 Dedham HLA Values of Dedham HLA Alex Sheh, Scientist - American Red Cross HLA services, Dedham, MA alex.sheh@redcross.org Provide the best service for our partners regionally and nationally Accurate Parallel testing for confirmatory typing Cost-efficient Non-profit organization Fast 3-day TAT, i.e. samples received on Monday are sent out on Wednesday. ASHI 9/29/2015

34 Collaboration with Omixon Started working with Omixon in late 2014 Robustness of the assay not as individualized attention to samples Multiplexing is key for our workflow software and reagents have been great. Receptive to feedback on software and assay Provided timely improvements to reduce bench time and reagents used Omixon team has been very supportive of our needs Supported our efforts to use shorter sequencing reads to reduce time and provided support on sample analysis Supported our efforts to adapt a Janus with 96-channel pipetting head for post-pcr

35 Collaboration with Omixon Number of alleles Total Samples 4216 Samples with SBT/SSO reference data 2094 Fully concordant with reference 2024 Discordant 70 Discordance level 3.34% Concordance level 96.66%

36 Early Access Program - Update Main changes to Holotype HLA during EAP: 2 more loci added Full pooling of all loci by default HLA Twin Software updated, new features added HLA Twin new features Locus-based subsampling HML and PDF export Whole gene consensus

37 Dimitri Monos

38 Tim Hague, Omixon

39 EAP Results Most of the labs involved in the EAP had never seen the Holotype HLA protocol before A number of labs had never done sequencing before All of the labs involved have been asked to use known samples, with known HLA types The results are for 5 and 7 locus data Most of the results are from fully pooled data

40 EAP Results 24 labs, 2530 samples, alleles 8788 alleles with known typings 1.7% background FTA (failure to amplify) rate Locus Total alleles (with known types)* HLA-A HLA-B HLA-C HLA-DQB HLA-DRB * = some known types were missing Unique alleles

41 Concordance Category Count Percent Alleles with known typings % Total discordances % Incorrect known typing % Discordant with QC failure 85 1% Discordance due to novelty % Still under investigation % Concordance = 2 field concordance

42 Ambiguity Number of loci HLA-A HLA-B HLA-C HLA-DQB1 HLA-DRB1 Unambiguous Cis/trans phase ambiguity th field ambiguity rd field ambiguity a 2 nd field ambiguity a 1 st field ambiguity Only a subset of data is available for ambiguity a = alleles differ in off-target exon sequence

43 Result Statistics Locus HLA-A HLA-B HLA-C HLA-DQB1 HLA-DRB1 Typed alleles Sensitivity 98.65% 97.48% 98.18% 95.71% 96.41% Specificity 99.98% 99.97% 99.97% 99.88% 99.96% PPV 98.66% 97.57% 98.18% 95.71% 97.04% NPV 99.98% 99.97% 99.97% 99.88% 99.95% TCC (Accuracy) 99.96% 99.95% 99.94% 99.76% 99.91% PPV = Positive Predictive Value NPV = Negative Predictive Value TCC = Type Correctly Classified (TP + TN / N) = Accuracy

44 Holotype HLA Next Steps DRB3/4/5, DPA1 in manufacturing European CE mark More optimizations to the assay protocol More optimizations to the software Automation Hamilton Starlet Perkin Elmer Janus 192 samples per MiSeq run

45 Announcements Competitor Special Free HLA Twin analysis for all evaluations of competing kit products No strings attached

46 The Omixon Website

47 The Omixon Academy

48 Questions? +1 (617)

49 LR-PCR Amplification Strategy HLA-A, B, C, and DQA1 (Example is HLA-B) Full gene characterization 5 UTR UTR HLA-DQB1 5 UTR UTR HLA-DRB1 Key region characterization 5 UTR UTR HLA-DPB1 5 UTR UTR

50 Effect of Varying Insert Sizes among Paired Reads Read Length 250 bases Read Length 250 bases Overlap of 100 bases Fragment Length 400 bp Read Length 250 bases Read Length 250 bases No Overlap Fragment Length 500 bp Read Length 250 bases Read Length 250 bases No sequence for 100 bases Fragment Length 600 bp

51 Optimized Size Selection Fragment Size Number of read pairs referring to fragments within the specified length interval

52 Long Distance Phasing B*07:02:01 T T AG T A C T B*41:01 C C T C C G T G

53 Allelic Balance The optimized assay design achieves allelic balance across all loci for accurate HLA allele determination without bias towards easy-to-sequence genomic regions.

54 Even Coverage Across a locus Across loci within a sample Between samples

55 Omixon HLA Twin (Software) Preconfigured Holotype HLA-specific settings Automated genotyping after MiSeq run Two algorithms for determining HLA genotypes Consensus Genotyping (Assembly) Statistical Genotyping (Alignment to IMGT/HLA) Traffic light system for data interpretation and workflow management PASSED WARNING FAILED

56 The Traffic Light System

57 Benefits of Holotype HLA Assay LR-PCR for fully characterized loci Easy Library Preparation Variable-insert paired-end sequencing for phasing High-throughput NGS on a MiSeq Unambiguous genotyping No reflexive testing

58 Benefits of HLA Twin Software Accurate genotyping with two orthogonal algorithms No manual intervention for good quality data Informative quality control metrics for confidence in genotyping results Traffic Light System for easy interpretation Novel allele detection Null alleles always resolved (splice variant and intronic) Export genotypes or consensus sequence for reporting

59 Holotype HLA vs Competitors Competing Products Occasional allele dropout Holotype HLA Very rare allele dropout 6+ hours hands on time About 4 hours hands on time (24 samples) Up to 24 samples per run Uneven coverage, dips in coverage at key exons Allele imbalance One algorithm, requires technician for every allele call High rates of ambiguity Up to 96 samples per run Deep & even coverage across every exon for each locus Balanced coverage depth across loci; and for both alleles at each locus Fully automated genotyping with two algorithms 0.2% ambiguity (based on 5 loci)

60 Whole Gene Consensus HLA Twin generates consensus sequences for both alleles across the whole region covered Independent of IMGT/HLA database Whole gene for HLA-A, B, C, DQB1, DQA1 Complete picture Novel allele detection Null allele detection Future proof

61 Whole Gene Consensus Example

62 The Other Allele