LaFranzo: They've done all four (AJ trio plus Asian son) as #GIAB FFPE. Has developed ctDNA standards with fragmented DNA.
5:20pm January 28th 2016 via Hootsuite
LaFranzo: Are working to develop their second-generation formalin-compromised product (to make it more realistic) #GIAB
5:14pm January 28th 2016 via Hootsuite
LaFranzo: But the mut allele freq differences between 0.04%, 0.15%. Question / discussion around age / quality of formalin #GIAB
5:12pm January 28th 2016 via Hootsuite
LaFranzo: How does formalin affect the performance of an assay? Affect mutation freq detection? Showed data of differences... #GIAB
5:08pm January 28th 2016 via Hootsuite
Natalie LaFranzo (Horizon Discovery) Asks if formalin-fixation 'how close to real-world can we produce?' A problem for a commercial co #GIAB
5:06pm January 28th 2016 via Hootsuite
Walgreens Pulls Back From Theranos - WSJ https://t.co/exGzr8Y03W
3:39pm January 28th 2016 via Hootsuite
Salit: Would a standard Minimum Information about Coriell needed? Lincoln, Jared from Counsyl both say 'yes'. Coriell: complicated. #GIAB
3:19pm January 28th 2016 via Hootsuite
Q: Does Coriell consent allow 'pull-back' like the HapMap consent? A: Yes, a handful have exercised it over 10y. #GIAB
3:08pm January 28th 2016 via Hootsuite
Lincoln: Points out the minor overlap between germline and somatic mutations that are critical variants. #GIAB
3:06pm January 28th 2016 via Hootsuite
Deficiencies at Theranos ‘pose immediate jeopardy to patient health’ - The Washington Post https://t.co/p8PzPrOWWr
3:02pm January 28th 2016 via Hootsuite
Salit: For accrual - the need for clear consenting to avoid ambiguity. Coriell (Nahid Turan) will work on expanding their consent #GIAB
3:01pm January 28th 2016 via Hootsuite
Intermountain Health: A concern about tumor calls, 'is it correct?' FFPE - concern over how close it is to a patient sample #GIAB
2:55pm January 28th 2016 via Hootsuite
Salit: Coriell as a source for clinical labs: is there a concern? Lincoln: Concern for vendors (whether TMO, ILMN, Horizon, SeraCare) #GIAB
2:53pm January 28th 2016 via Hootsuite
Lincoln (In Vitae) - gene panels cost about $500-700 (fully-loaded). Q:Will a company pay for it gratis? A:Some but not a lot. #GIAB
2:46pm January 28th 2016 via Hootsuite
Q: Cost of lymphoblastoid line creation? About $700-800. Can Karyotype for <$1K additional #GIAB
2:44pm January 28th 2016 via Hootsuite
Kalman: Working out how large a pilot project this will become (funding dependent). PII: to 96 lines, add new (created) lines #GIAB
2:43pm January 28th 2016 via Hootsuite
Kalman: Their pilot: 96 existing Coriell lines; sequence genes using variety of NGS assays (WES, panels, WGS); confirm via Sanger #GIAB
2:41pm January 28th 2016 via Hootsuite
Kalman: Synthetic lines or plasmid spike-ins as well. "Or other things as people come up with them" #GIAB
2:40pm January 28th 2016 via Hootsuite
Kalman: Perhaps 100's of samples; Mendelian disorders, cardiomyopathy, inh cancer. Use existing Coriell lines; create new lines... #GIAB
2:39pm January 28th 2016 via Hootsuite
Kalman: Need many samples per gene to det sens/spec; and many variant types. Want assay/tech comparision, standardization #GIAB
2:37pm January 28th 2016 via Hootsuite
Kalman: #GIAB is 'just a couple of genomes; 'ideally, labs should validate assay FOR EACH GENE interrogated!'
2:36pm January 28th 2016 via Hootsuite
#GIAB breakout on reference materials: Lisa Kalman (CDC) for a GeT-RM update (Genetic Testing Reference Materials) https://t.co/gBnkcQ9Jrf
2:35pm January 28th 2016 via Hootsuite
Q: Idea for implementation model? Zook: Likely web-based to upload files, so details unclear. "Anyone wanting to participate welcome" #GIAB
1:28pm January 28th 2016 via Hootsuite
Q: Any timeline for usability? Zook: PrecisionFDA is working on improving their simple comparison tool 'within a year' #GIAB
1:27pm January 28th 2016 via Hootsuite
Jacobs: Also may be an issue of funding as well; PrecisionFDA may help here. 'We can't report a "bland marginalized number"' #GIAB
1:24pm January 28th 2016 via Hootsuite
Steve Lincoln (In Vitae):Q:How far will the tool be accessable? Jacobs: A UI issue - the barrier is implementing web-delivery to users #GIAB
1:23pm January 28th 2016 via Hootsuite
Jacobs: Are two alternative HGVS var's equivalent, or not? Best-case performance of a tool, and the worst case? #GIAB
1:20pm January 28th 2016 via Hootsuite
Jacobs: His tool models uncertainty; propagating no-call throughout the pipeline. Illumina's tool on BaseSpace, his experimental #GIAB
1:19pm January 28th 2016 via Hootsuite
Kevin Jacobs (23andMe): "Imagine talking about the semantics of VCF files over and over again." Comparisons, straightening out formats #GIAB
Q: Def'n of 'small variant'? Zook: it's tool-dependent. Most are <25bp #GIAB
1:18pm January 28th 2016 via Hootsuite
Salit: They want to re-package var calling for NA12878, perhaps with DNA Nexus. Hope to have a transparent, usable tool #GIAB
1:13pm January 28th 2016 via Hootsuite
Zook: Acknowledges Liz Mansfield (FDA), new members, long list for the #GIAB steering committee.
1:11pm January 28th 2016 via Hootsuite
Zook: There are multiple dimensions of accuracy: predicted SV existence, SV type, size, breakpoints, and exact sequence. #GIAB
1:09pm January 28th 2016 via Hootsuite
Zook: Also hard w/complex variants near boundaries of high-confidence regions. "Always manually inspect a subset of FPs/FNs" #GIAB
1:06pm January 28th 2016 via Hootsuite
Zook: Benchmarking problem: many analytes (variants and non-variants); easy to benchmark high-confidence (but easy to call) #GIAB
1:05pm January 28th 2016 via Hootsuite
Zook: Shows architecture (vial Peter Krusche on their github). Truth vs Query VCFs are compared, quantified, and counted #GIAB
1:03pm January 28th 2016 via Hootsuite
Zook:Proposing TP/FP/FN/TN in 4 ways: loose, allele, phenotype, phasing matches. 'TN have an infinite # of potential alleles' #GIAB
1:02pm January 28th 2016 via Hootsuite
Zook: Github also has discussion, want to write this up, vcfeval (Len Trigg) hap.py (Peter Krusche) also have tools. #GIAB
1:00pm January 28th 2016 via Hootsuite
Zook: Points out tool vgraph Kevin Jacobs (@23andme), also simulating reads with non-SNP ClinVar variants to demo importance of tools #GIAB
12:59pm January 28th 2016 via Hootsuite
Zook: Points out many publications using NA12878 #GIAB Points out Github repo https://t.co/gv4JZlBDyH for benchmarking tools
12:57pm January 28th 2016 via Hootsuite
Zook: Working with DNA Nexus and Parliament alg. for SV; also MetaSV, svclassify. Testing best-practices for manual inspection #GIAB
12:52pm January 28th 2016 via Hootsuite
Zook: Shows data from this Aug '15 Nature Methods https://t.co/VDkfXqfBES using @pacbio on NA12878 #GIAB
12:51pm January 28th 2016 via Hootsuite
Zook: For SNP/indel integration - implementing 'refined integration methods'; much easier to replicate. #GIAB
12:49pm January 28th 2016 via Hootsuite
Zook: Some groups doing methylation analyses (unintended), but with @Pacbio data has people working on it #GIAB
12:46pm January 28th 2016 via Hootsuite
RT @mason_lab: #GIAB all data as of January 2016 - 12 methods for WGS! Fantastic! ftp://ftp-trace.ncbi.nih.gov/giab/ftp/release https://t…
12:45pm January 28th 2016 via Twitter Web Client
Zook: Long range phasing (10X and Moleculo) not incorporated yet. SV candidate calls generated by >15 groups, >20 algorithms #GIAB
12:44pm January 28th 2016 via Hootsuite
Justin Zook: Reviews participants and updates on #GIAB analysis group, still recruiting to assist with the work.
12:43pm January 28th 2016 via Hootsuite
Salit: Four microbial genomes, integrated with both long- and short-read data. #GIAB Some slideshare material: https://t.co/ei0HnyeBSO
12:41pm January 28th 2016 via Hootsuite
Salit:Microbial gDNA: salmonella, staph, pseudomonas, clostridium. (Slide from Nate Olsen) GC from 29 - 67% sold as a set #GIAB
12:36pm January 28th 2016 via Hootsuite
Salit: SEQC2: includes WGS and targeted. Platforms, samples. RM8398 is CEPH daughter; slides of AJ trio, Asian son but not parents #GIAB
12:35pm January 28th 2016 via Hootsuite
