Santani: Missed by first lab - b/c OMIM didn't have a disease name. First report of SOX5 point mutation same as large deletion #NGDx15
9:58am August 19th 2015 via Hootsuite
Santani:Now covers these two cases; classic dx odyssey. SOX5 - no disease associated with SOX5 del. 18 cases in lit #NGDx15
Santani: Of the rest - needed more phenotype data from physician. 2 prior were negative, now ID'd two new molecular diagnosis #NGDx15
9:57am August 19th 2015 via Hootsuite
Santani: But the clinical validation piece - variant filtration and interpretation process. 20 samples - 16/20 concordant #NGDx15
9:56am August 19th 2015 via Hootsuite
Santani:Their validation; used GIAB 'gold standard' NA12878, terrific way to test sens and spec of WES. #NGDx15
9:55am August 19th 2015 via Hootsuite
Santani: Their LIMS will generate reports as well.Rules for compliance. #NGDx15
9:53am August 19th 2015 via Hootsuite
Santani: The KB part of LIMS - not only tracking samples, but also genomic-savvy. 2y of effort: phenotype collection, images, test #NGDx15
Santani: "We will not use Excel" - 100's of thousands of vars, over 10 people needing to access data. Wanted a knowledge base #NGDx15
9:52am August 19th 2015 via Hootsuite
Santani:The final few - from 3 to 40 variants evaluated for pathogenicity. ACMG classifications 'outstanding' http://t.co/9LFZGOqHpJ #NGDx15
9:51am August 19th 2015 via Hootsuite
Santani: Each step - multiple databases, an intensive process. 100's of exome cases/mo. Trio WES - many families from Middle East #NGDx15
9:49am August 19th 2015 via Hootsuite
Santani:1,141 by mutation type; of that 512 disease genes; then 40 clinician board review select. Then 3 de novo selected #NGDx15
9:48am August 19th 2015 via Hootsuite
Santani: Many questions, each with a report. GATK calling, but filtering. Quality var, rare, mutation type, disease genes #NGDx15
9:47am August 19th 2015 via Hootsuite
Santani: Showed complex pipeline for their CAP compliant bifx pipeline. Reports,overall quality mgmt, low coverage. #NGDx15
9:46am August 19th 2015 via Hootsuite
Santani: Informatics - a challenge. 'It was like a foreign language exchange program' SAM? CAP? #NGDx15
9:44am August 19th 2015 via Hootsuite
Santani: On a V5 now, at 98% coverage at >20x. They had about 100 panels for their customers to accommodate. #NGDx15
9:43am August 19th 2015 via Hootsuite
Santani: Collaborated with Emory (M Hegde) and Partners (B Funke), 3y effort: targeted the low-coverage regions to supplement WES #NGDx15
9:42am August 19th 2015 via Hootsuite
Santani: Medical exome def'n: 5,000 genes of medical significance. Then sliced further into panels (e.g. cardiomyopathy) #NGDx15
9:41am August 19th 2015 via Hootsuite
Santani:Benchmarking - whole exome, medical exome, cardiomyopathy panel. Bases <20x, fully covered exon (>20x) metrics.75%/82%/85% #NG
9:40am August 19th 2015 via Hootsuite
Santani:Director, technical supervisor, data analysts, genetic counselors etc. An ongoing challenge - attracting top talent NGDx15
9:39am August 19th 2015 via Hootsuite
Santani: Request for PhD-level technologists: new job descriptions needed. Bioinformatics specialist is not a software developer. #NGDx15
9:38am August 19th 2015 via Hootsuite
Santani: There were no compliance checklist 7 years ago. Informed consent, spent time determining return of results #NGDx15
9:37am August 19th 2015 via Hootsuite
Santani: Surveyed needs across clinics and prioritized tests to demonstrate need. Also acquired an R&D budget. Composed compliance #NGDx
9:36am August 19th 2015 via Hootsuite
Santani:Has many platforms, both NGS and microarray. Composed a business plan with ROI - why important and who was it for. #NGDx15
9:35am August 19th 2015 via Hootsuite
Santani: Has been working on this for the 5-7 years. Working on inherited diseases, with pediatric setting in mind #NGDx15
9:33am August 19th 2015 via Hootsuite
Avni Santani (UPenn and Children's Hosp of Philadelphia) "Implementation of clinical exome sequencing" #NGDx15
9:32am August 19th 2015 via Hootsuite
Daber: Larger panels 310 genes, 4K bi-allelic SNPs, address fusions, inversions, CNV, merge DNA and RNA variants #NGDx15
9:28am August 19th 2015 via Hootsuite
Daber:Their metrics: 76% of samples have 1 or more actionable mutation; 1.3 per 0.16 VUS. <0.1% assay failure rate #NGDx15
9:27am August 19th 2015 via Hootsuite
Daber: ctDNA is fragmented, small input, may use same process for that in their development work #NGDx15
9:26am August 19th 2015 via Hootsuite
Daber: Shows sensitivity down to 4% at 20ng input. 500pg input down to about 20% #NGDx15
9:25am August 19th 2015 via Hootsuite
Daber: They can go down to 100pg (15-18 cells) but at cost of primer-dimer. Asks if they could construct from collection of few CTCs #NGDx15
9:24am August 19th 2015 via Hootsuite
Daber: FFPE tumor DNA varies in quality widely - need to optimize library prep for small fragments and low input (120bp, 1ng) #NGDx15
9:23am August 19th 2015 via Hootsuite
Daber: Showed pie chart of NSCLC genes. (Quoted the '85%' number from Foundation paper, a bit of a red herring there.) #NGDx15
9:22am August 19th 2015 via Hootsuite
Daber: Criteria - clinical utilty, cost effectiveness, fast turnaround, reduce 'QNS' rates #NGDx15
9:21am August 19th 2015 via Hootsuite
Daber: Selecting content? Listed @iontorrent @illumina RainDance, MS, Foundation One, MATCH - only 109 common from 486 genes#NGDx15
9:20am August 19th 2015 via Hootsuite
Daber: Gene fusions - PCR using RNA. DNA: "Brute force tiling'. They use Archer, semi-amplicon based (one-side known, other unk) #NGDx15
9:19am August 19th 2015 via Hootsuite
Daber: Even so with intermediate dels: bioinformatics has to accomodate - primer cross-talk, looking at read-pairs. #NGDx15
9:16am August 19th 2015 via Hootsuite
Daber: AmpliSeq two different primers in same pool; RainDance amplicons within same drop; TSCA same DNA strand. Archer - no problem #NGDx15
9:15am August 19th 2015 via Hootsuite
Daber:What about larger deletions? Hyb-based will work better than PCR-based; you can but takes effort - 'it's complicated' #NGDx15
9:14am August 19th 2015 via Hootsuite
Daber: Homologous regions are eliminated in order to improve sensitivity / specificity. #NGDx15
9:13am August 19th 2015 via Hootsuite
Robert Daber (BioReference) "Development and implementation of clinical NGS testing: assay development and informatic challenges" #NGDx15
9:12am August 19th 2015 via Hootsuite
Fernandez:Second case of intrahepatic cholangiocarcinoma, had IDH1, common according to #TCGA. In clinical trial, 'pt doing well' #NGDx15
8:59am August 19th 2015 via Hootsuite
Fernandez: Concludes with a papillary thyroid case report: sample had both BRAF and NRAS mutations. Debate regarding classification #NGDx15
8:57am August 19th 2015 via Hootsuite
Fernandez: Variants via cytology vs resection smpls: concordance of 13 samples, 92%. 1/13 cellularity low #NGDx15
8:54am August 19th 2015 via Hootsuite
Fernandez: Same looked at via NGS: 29% cytology, # of indeterminates go down to 3% from 8% prior. Surgical: 27% about same #NGDx15
8:53am August 19th 2015 via Hootsuite
Fernandez: Comparing real-time PCR results for both cytology and surgical samples: 22% vs 31% mutation (more material) #NGDx15
8:52am August 19th 2015 via Hootsuite
Fernandez: Cytollogy vs surgical resection for lung ca: ave neoplastic cellularity of FNA was 23%, sig lower than resections 67% #NGDx15
8:50am August 19th 2015 via Hootsuite
Fernandez: Showed pie chart of NSCLC EGFR 26%, KRAS 28%, BRAF 3%. Has made a difference in clinical care. #NGDx15
8:49am August 19th 2015 via Hootsuite
Fernandez: Their costs for 450 genes about $250. #NGDx15
8:47am August 19th 2015 via Hootsuite
Helen Fernandez (Weill Cornell NY) "Utility of implementing clinical NGS assays as standard of care in oncology" #NGDx15
8:44am August 19th 2015 via Hootsuite
RT @softwareguru1: Introducing the single-cell maze runner http://t.co/2VefGK2bKy via @physorg_com
8:01am August 19th 2015 via Hootsuite
