Mittelman: Did you miss a gene or exon in your panel? There are dropouts - unacceptable in clinical setting. QC steps passed along #TCGC15
7:33pm June 23rd 2015 via Hootsuite
Mittelman: Var callers depends on the application. BioIT launched; 200K on it. Acknowledges @gabeinformatics and @gholsonlyon #TCGC15
7:28pm June 23rd 2015 via Hootsuite
Mittelman: Same var caller (GapK) but Venn diagram showed - different variants. Worked with J Zook NIST Genome in a Bottle (GIAB) #TCGC15
7:26pm June 23rd 2015 via Hootsuite
Mittelman: Bowtie is good - but insensitive to quality. GCAT gives insight to other metrics. 'You may get different variants' #TCGC15
7:25pm June 23rd 2015 via Hootsuite
Mittelman: Can do read alignment comparisons, next several examples. Novoalign and BWmem both are best, but have variation. #TCGC15
7:24pm June 23rd 2015 via Hootsuite
Mittelman:Highnam '15 ref http://t.co/NTRCBWs7Mi pipeline free to use called GCAT here: http://t.co/p1UwqF8u1F catching best methds #TCGC15
7:23pm June 23rd 2015 via Hootsuite
Mittelman: Partnered with DNANexus (alignment and var calling) Tute does annotation and reporting. '12 ref http://t.co/FpWJMj0ufI #TCGC15
7:21pm June 23rd 2015 via Hootsuite
Mittelman: Curated datasources, private and other public ones, fast and scalable annotation. Prenatal to inherited disorders #TCGC15
7:18pm June 23rd 2015 via Hootsuite
David Mittelman (Tute Genomics) "Democratizing regulated clinical sequencing" #TCGC15
7:16pm June 23rd 2015 via Hootsuite
Fortney:Q:HLI's approach is big data - does clinical data help? A:It can help (healthy vs. not), not clear-cut how best to use it #TCGC15
6:42pm June 23rd 2015 via Hootsuite
Fortney: Used GWAS data from disease for overlap to longevity SNPs. Traits and disease GWAS listed, developed p-value weighting #TCGC15
6:28pm June 23rd 2015 via Hootsuite
Fortney:WGS via Complete Genomics; didn't find any replicated. Used PGP data, made data available '14 PLOS http://t.co/UjtGaaiuor #TCGC15
6:24pm June 23rd 2015 via Hootsuite
Fortney: Only 17 supercentarians alive in the US. Many highly functional, ave age of death 113. 2/17 had major age-rel disease #TCGC15
6:18pm June 23rd 2015 via Hootsuite
Fortney: Centenarian GWAS - only APOE - but low #'s (100's of samples). Two stories: rare var's w/17 >110yo (super-centenarians) #TCGC15
6:17pm June 23rd 2015 via Hootsuite
Fortney: One example of longevity - about 25%. Kahn family, four siblings all >100yo. They work for a long time. 83% less stroke etc #TCG
6:16pm June 23rd 2015 via Hootsuite
Kristen Fortney (Stanford Univ) "Whole-genome sequencing of the world's oldest people" #TCGC15
6:14pm June 23rd 2015 via Hootsuite
Kong: Concludes:(one of many): "analytical validity of current state-of-the-art algorithms is not satisfactory" #TCGC15
6:08pm June 23rd 2015 via Hootsuite
Kong: Comparison of VCFs from 1 FASTQ file (from GIAB) illustrated. Fig from '11 ref http://t.co/UOQSoCAj4j #TCGC15
6:04pm June 23rd 2015 via Hootsuite
Kong: Using @hail_cser and 100 WGS as part of a project called MedSeq '14 Genet Med ref: http://t.co/4xwcBXGWdJ Polygenic risk score #TCGC15
5:50pm June 23rd 2015 via Hootsuite
Kong: How can probabilistic predictions of disease be incorporated into personal and clinical decision-making? #TCGC15
5:48pm June 23rd 2015 via Hootsuite
Kong: What if we had WGS of millions of individuals? '13 ref http://t.co/gUswQ96U25 Actionable muts and for offspring #TCGC15
5:45pm June 23rd 2015 via Hootsuite
Kong: Mentions Mark Cuban's tweets about getting a quarterly blood baseline. Human exposome (Dean Jones) 400K chemicals #TCGC15
5:39pm June 23rd 2015 via Hootsuite
Kong: Eric Topol's 2014 Cell paper http://t.co/bHjl40SfN4 From prewomb to tomb: prevent disease, infectious disease #TCGC15
5:37pm June 23rd 2015 via Hootsuite
Kong: Near-term focus on cancers, longer term to improve healthcare. Snyder's personal 'omics '12 Cell ref http://t.co/Dt5Pxzafax #TCGC15
5:35pm June 23rd 2015 via Hootsuite
Sek Won Kong (Harvard Medical School and Boston Children's Hospital) "An approach to WGS as a lifelong health resource" #TCGC15
5:32pm June 23rd 2015 via Hootsuite
Garlick:Q:Curation of info, db's? A:Will be doing an integrated QC study, look at QC metrics to pinpoint where QC should take place #TCGC15
4:17pm June 23rd 2015 via Hootsuite
Garlick:Q:Stability? A:Patent method filed last month. Q:FFPE is simulated? A:Not planning cell lines, PDX instead. #TCGC15
4:16pm June 23rd 2015 via Hootsuite
Garlick:Q:Tools to FDA - how to correlate w/assay mfrs? A:SeraCare is agnostic to assays. Validation of product to intended use. #TCGC15
4:14pm June 23rd 2015 via Hootsuite
Garlick:Q:Shared data with Illumina? A:Yes the conversation will start #TCGC15
Garlick:Q:Cost an issue; how about spike-in control? A:Planning to do with ctDNA and plasma, not with FFPE tissues though #TCGC15
4:13pm June 23rd 2015 via Hootsuite
Garlick:Q:Brings up fetal aneuploidy paper; not intended use (i.e. non-pregnant sample for aneuploidy) A:Accepted - made the news #TCGC15
4:12pm June 23rd 2015 via Hootsuite
Garlick: Summary - LDT to IVD transition will require multiple QC tools. #TCGC15
4:10pm June 23rd 2015 via Hootsuite
Garlick: Maternal bckgnd, fetal fx from trophoblasts. DNA stabilized, put into plasma. T21 linearity shown. Commutable matrix #TCGC15
4:08pm June 23rd 2015 via Hootsuite
Garlick: Same principles to ctDNA, dev. non-invasive prenatal screening product (Aug). T13, T18 and T21, 4-8% fetal fx. 2-14% perf. #TCGC15
4:07pm June 23rd 2015 via Hootsuite
Garlick:Plans to do germline controls with same method. Can get multi-site proficiency. Developing NGS-char FFPE tissues #TCGC15
4:05pm June 23rd 2015 via Hootsuite
Garlick: Variability over time: Levey-Jennings plot, 28 runs over 8 mo: specific mutations. #TCGC15
4:04pm June 23rd 2015 via Hootsuite
Russell Garlick at #TCGC15 talking about @SeraCare and Levey-Jennings plots http://t.co/kMELnMOKjv
Garlick: Some assay differences between AmpliSeq cancer panel and TruSeq on the MiSeq. CV's at 5%. Shows linearity at 10%, 15% etc #TCGC15
4:01pm June 23rd 2015 via Hootsuite
Garlick: Goal to challenge seq and pipelines. Performance of two assays - hyb-based vs. AmpliSeq, shows PCR dropout #TCGC15
3:59pm June 23rd 2015 via Hootsuite
Garlick: Constructs have been sequenced, quantitated via digital PCR, spiked into background of GIAB genomic DNA. Adjust allele freq #TCGC15
3:58pm June 23rd 2015 via Hootsuite
Garlick:Q:6-bp tag affects alignment? A:Can be specifically searched for and awareness of it helps analysis #TCGC15
3:57pm June 23rd 2015 via Hootsuite
Garlick: 6-bp engineered Internal Quality Marker included. Synthetic target in wild-type background gDNA. Showed del next to 6bp #TCGC15
3:55pm June 23rd 2015 via Hootsuite
Garlick: Seraseq product developed based upon Frederick Nat'l Lab for Cancer Research (NCI), CRADA for co-development #TCGC15
3:53pm June 23rd 2015 via Hootsuite
Garlick: 'There will be assay drift over time' - sensitivity, accuracy, reprod, LOD needs to be assessed #TCGC15
3:52pm June 23rd 2015 via Hootsuite
Garlick: 12 QC metrics - lots of variability in front-end. For oncology assays - only 2y ago 50 genes; today 613. #TCGC15
3:51pm June 23rd 2015 via Hootsuite
Garlick:Intention is to take these materials to the FDA. One paper 6/8 labs got this test correct http://t.co/NAFLLs2i0x #TCGC15
3:49pm June 23rd 2015 via Hootsuite
Garlick: Pressures of cost - but what metrics matter? What drifts over time? W/o standards over time, an unknown. #TCGC15
3:47pm June 23rd 2015 via Hootsuite
Garlick: NGS transition to clinical assays - from 40 to perhaps 250 labs in 2y: targeted gene panels, some WES, fusion RNA #TCGC15
3:46pm June 23rd 2015 via Hootsuite
Garlick: Oncology, biosynthetic details, some data from PDX (patient-derived xenografts), #TCGC15
Garlick:5-10% allele freq, ID driver genes. Reagent changes, pipeline changes, instrument changes - all need new controls #TCGC15
3:44pm June 23rd 2015 via Hootsuite
