Implementation & development of NGS in the diagnostic lab

Transcription

1 Implementation & development of NGS in the diagnostic lab Helen Lindsay Clinical Scientist Yorkshire Regional DNA Laboratory Deauville 16/12/11

2 NGS in the Leeds DNA lab The first NHS lab to implement NGS as a routine service Illumina GAIIx platform single 80bp reads Instrument run once a fortnight; each run is shared between University and NHS Target enrichment achieved by long PCR Multiple samples run per lane through use of tagged adaptors

3 Validation Parallel study 53 patients tested in parallel with conventional BRCA panels 437 heterozygous variants 106 homozygous variants 41 distinct variants 100% concordance Morgan et al (2010); Human Mutation 31(4):484 analysis of TP53, BRCA1 and BRCA2 mutations in 55 patients previously screened by conventional sequencing 604 variants detected 100% concordance Test sensitivity calculated by considering the number of unique variants detected in validation studies 95% confidence that false negative rate is <5% Disease-specific validation for new services Smaller scale validation panel of known variants Primer design to reduce risk of allelic-dropout SNP monitoring

4 NGS services offered Service Breast Cancer Lynch syndrome (HNPCC) Pheochromocytoma Marfan syndrome Hypertrophic Cardiomyopathy Genes BRCA1 BRCA2 with MLPA MSH2 MLH1 MSH6 with MLPA RET, VHL, SDHB, SDHC, SDHD, SDH5, PRKARIA, TMEM127 with MLPA FBN1 with MLPA MYH7 MYBRC3 TNNT2 TNNI3 Date launched February 2010 October 2010 May 2011 June 2011 June 2011 Number of NGS reports issued as of Oct 2011 Cost (NHS referrals) Comments Transferred from existing service PMS2 testing also available; Transferred from existing service Restricted analysis for RET NEW service Complementary TGFBR1/ TGFBR2 service for Loeys- Dietz syndrome in development NEW service NEW service

5 Parallel workflows Service 1 Service 2 Service 3 Standard PCR (where required) LR PCR LR PCR LR PCR Library Preparation Standard PCR (where required) Clonal Sequencing Sanger Sequencing Data Analysis and Interpretation Reports issued Reports issued Reports issued

6 Library preparation I Target enrichment long PCR of the complete CDS Clean-up and quantification of PCR products Shearing of pooled PCR products using the Covaris S2 instrument; verification on the Agilent Bioanalyzer

7 Library preparation II Beckman SPRIworks Fragment Library system library prep of 10 samples takes 5 hours (~10 minutes hands-on) Enrichment PCR; verification on Agilent Bioanalyzer Quantification and pooling of tagged samples > Illumina cluster generation and sequencing on cbot/gaiix

8 Illumina cluster generation DNA hybridised to flow cell Fragments clonally amplified Sequencing primer hybridised

9 Sequencing Copy template using fluorescently labelled, reversibly terminated nucleotides Capture fluorescence signal after each incorporation step Remove blocking group to allow addition of next base

10 Data analysis NextGENe (SoftGenetics) software conversion of qseq to fasta removal of low quality files tag sorting alignment to reference sequence(s) mutation detection read depth information customised spreadsheets for assisted analysis

11 mutation report alignment coverage

12 MLH1 c.1852_1853delaainsgc mutation report alignment coverage

13 mutation report alignment coverage

14 BRCA1 c.1175_1214del40 mutation report alignment coverage

15 Data analysis Generation of mutation reports manual inspection of sequence data if necessary to confirm variants and exclude artefacts Generation of read depth information custom spreadsheets automatically check depth at each base across the region of interest a minimum threshold of 50-fold read depth is applied for diagnostic screening

16 Mutation report

17 Read depth

18 Post analysis Confirmation of all sequence variants other than known neutral polymorphisms by Sanger sequencing (under review) Sanger sequencing of any regions with less than 50-fold read depth Creation of a diagnostic report combining sequence data with MLPA results

19 Discussion To date we have sequenced and reported >1000 patients All mutations and UVs confirmed by Sanger sequencing 40% reduction in average test cost 50% reduction in hands-on staff time 30-40% reduction in consumables costs Process is CPA accredited Participated in 2010/11 NEQAS schemes for BRCA1/2

20 Impact of NGS Turnaround times Established NGS services are reported comfortably within Government targets Feedback from clinicians indicates that results may be received in time to influence treatment options (e.g. prophylactic surgery) Comprehensive tests The pheochromocytoma service represents a shift in testing, with parallel testing of 8 relevant genes replacing former testing of the same genes sequentially

21 Impact of NGS A mainstream technique Over half the molecular workload in Leeds is now using NGS Reduced lab costs ~40% reduction in test cost Increased capacity Increased reliability Close working relationship with research groups Easily accessible & auditable sequence data from 100s of patients NGS referral origin by quarter Overseas/Private UK Yorkshire/Northern 50 0 Q1 Q2 Q3 Q4 Q5 Q6 Q7

22 Challenges New technology no best practice guidelines validation SOPs Long PCR design GC rich regions allele drop-out Assay limitations large deletions homopolymer regions Training new techniques both lab work and data analysis Information technology bioinformatics data storage New services confirmation of NGS results Interpretation of multi-gene tests UV analysis Maximising run capacity

23 BRCA review BRCA1 & BRCA2 service launch March 2010 NGS sequence analysis MLPA Over 800 reports issued BRCA1 67 pathogenic (9 MLPA) 48 unclear pathogenicity BRCA2 64 pathogenic (2 MLPA) 68 unclear pathogenicity

24 HNPCC review MLH1, MSH2, MSH6 service launch October 2010 NGS sequence analysis MLPA Over 150 reports issued 5 MLPA pathogenic variants MLH1 7 pathogenic variants 8 unclear pathogenicity MSH2 12 pathogenic variants 5 unclear pathogenicity MSH6 2 pathogenic variants 8 unclear pathogenicity PMS2 service (not NGS) APC service (transfer to NGS in progress)

25 Pheochromocytoma review Service launched June 2011 Parallel testing of eight genes 69 reports to date 9 pathogenic variants 6 SDHB 3 SDHD 4 variants of unclear pathogenicity 1 RET 2 SDHD 1 SDHAF2

26 Array capture 20 kb Database of Genomic Variants: Structural Variation (CNV, Inversion, In/del) Segmental Dups Repeating Elements by RepeatMasker RepeatMasker Gene_List BRCA1 My Custom Track Switches from over 20 PCRs to get two genes to one PCR to enrich many genes (pathway or syndrome-driven testing) BC-CC_exons_RM _ 3002 forward_reads forward_reads _0 _ 2848 reverse_reads reverse_reads _0 RefSeq Genes Scale chr17: DGV Struct Var Scale chr13: DGV Struct Var Segmental Dups RepeatMasker BRCA2 BC-CC_exons_RM 2866 _ Segmental Dups 20 kb Database of Genomic Variants: Structural Variation (CNV, Inversion, In/del) RepeatMasker TP53 TP53 Duplications of >1000 Bases of Non-RepeatMasked Sequence BC-CC_exons_RM 2033 _ Repeating Elements by RepeatMasker GC Percent in 5-Base Windows 2 kb Database of Genomic Variants: Structural Variation (CNV, Inversion, In/del) Duplications of >1000 Bases of Non-RepeatMasked Sequence Repeating Elements by RepeatMasker Gene_List My Custom Track forward_reads Gene_List My Custom Track forward_reads forward_reads forward_reads 0_ 1983 _ 0_ 2483 _ reverse_reads reverse_reads reverse_reads reverse_reads 0_ 0_ BRCA2 GC Percent RefSeq Genes GC Percent in 5-Base Windows Scale :chr17 DGV Struct Var Duplications of >1000 Bases of Non-RepeatMasked Sequence TP53 TP53 TP53 TP53 TP53 TP53 TP53 RefSeq Genes GC Percent in 5-Base Windows GC Percent BRCA1 BRCA1 BRCA1 BRCA1 BRCA1 BRCA GC Percent

27 Further developments Alternative targeting strategies Larger targets (e.g. Agilent SureSelect) cardiome 54 genes, 1297 exons, 0.6Mb cancer chip / retinome / kinome / ciliome etc. autozygome 10 genes identified in Leeds using custom arrays Whole exome/ whole genome

28 Further developments Risk assessment of NGS process Can Sanger sequencing of all clinically significant variants detected by NGS be abandoned? Further library preparation automation Streamlined variant assessment and classification

29 Further developments Instruments Faster instrument for rapid turnaround (MiSeq) Predictive tests Where clinically useful for result sooner than 40 working days Quality at least comparable to GAIIx Higher output instrument for economic sequencing of larger targets (HiSeq) Large targets / whole exome / genome Genome-wide copy number analysis Streamlined variant assessment and classification Required for larger targets

30 NGS services in development Li Fraumeni syndrome FAP Optic Atrophy FEVR Long QT syndrome Loeys-Dietz syndrome Aicardi- Goutieres syndrome Epilepsy - cerebral malformations Epileptic Encephalopathy TP53 APC OPA1 FZD4 LRP5 NDP TSPAN12 KCNQ1 KCNH2 KCNE1 KCNE2 SCN5A TGFBR1 TGFBR2 TREX1 RNASEH2B RNASEH2C RNASEH2A SAMHD1 ARX DCX LIS1 GPR56 TUBA1A TUBB2B SCN1A CDKL5 SLC2A1 (GLUT1)

31 Acknowledgements Leeds DNA lab Nick Camm Rachel Robinson Ruth Charlton David Cockburn Leeds Institute of Molecular Medicine Jo Morgan Graham Taylor