EM: Will open up CIViC crowdsourcing at #ASCO15 later this year. Var & evidence statements included. #AGBT15
10:02am February 26th 2015 via Hootsuite
EM: CIViC database http://t.co/xIQTqYdx8k a clinical data curation of cancer variants. Open-source, crowd-sourced info sharing #AGBT15
EM:Sometimes predominant mut's missed; DoCM website is a KB-based approach to the cancer genome http://t.co/HEqgcobNIa #AGBT15
10:00am February 26th 2015 via Hootsuite
EM: 2nd ref: Ramikissoon et al '13 http://t.co/tSNf2LvXBP #AGBT15
9:59am February 26th 2015 via Hootsuite
EM: Now onto pediatric low-grade glioma. Example of non-resp to chemo; ~500 somatic SNVs. Themes Zhang '13 http://t.co/eGdfNF3P9z #AGBT15
9:58am February 26th 2015 via Hootsuite
EM: Looking at 0d & 30d retrospectively (with long-term followup): everyone is highly different. #AGBT15
9:55am February 26th 2015 via Hootsuite
EM: Proposing WES, augmented with RMG probes for 259 AML genes and transcriptome sequencing. Blog post: http://t.co/dcYdE7SGal #AGBT15
9:53am February 26th 2015 via Hootsuite
EM: Adding genomics to AML - 200 samples contrib to TCGA published in NEJM 2013 here http://t.co/tUaiEJuuYH #AGBT15
9:48am February 26th 2015 via Hootsuite
EM: Extraordinary filtering to make up for absence of matched normals; mutation rates recapitulate TCGA results PIK3CA, MLL3 etc #AGBT15
9:45am February 26th 2015 via Hootsuite
EM: Ave 25yo FFPE 'a tragedy' due to degradation of nucl acids. 50-300ng input; 83 gene panel probes; 625/632 passed (80% >20x) #AGBT15
9:44am February 26th 2015 via Hootsuite
EM:For estrogen receptor + br ca: acknowledges @susangkomen grant. Nielsen 2010 http://t.co/NHq6faCH0m #AGBT15
9:42am February 26th 2015 via Hootsuite
EM: Will touch on 4 areas: markers of late relapse ER+ disease, translating AML, challenges to transition... #AGBT15
9:40am February 26th 2015 via Hootsuite
Elaine Mardis, Washington University School of Medicine. “Technical and Informatics-Based Developments in Cancer Genomics” #AGBT15
9:39am February 26th 2015 via Hootsuite
EE:A2:You have to get outside the RL limitation, why BAC-based sequencing is used. #AGBT15
9:37am February 26th 2015 via Hootsuite
EE:Q:Optical mapping, UPenn obs. to 70kb A:A small proportion of total events; repeat structure longest RL the limitation #AGBT15
9:36am February 26th 2015 via Hootsuite
RT @KMS_Meltzy: It all comes back to the BACs #AGBT15
9:35am February 26th 2015 via Hootsuite
EE:Q:Consensus reads? A:Not SMRTbells, after Quiver 99.99% Qv 40 to 50, Cp to Sanger as well as BAC sequencing #AGBT15
EE:Q:Implications for Dx? A:100's in coding seq; largely untapped space to explore #AGBT15
9:34am February 26th 2015 via Hootsuite
EE:Q:Upper limit of detection size? A:Insertions is 50bp-5kb is novel; Phase III data not included. Tandem dupl: to 40kb due to RL #AGBT15
9:33am February 26th 2015 via Hootsuite
EE: Shoutout to Jason Chin of @PacBio. Q:FP/FN rate of genotyping? A:No metrics yet #AGBT15
9:31am February 26th 2015 via Hootsuite
EE:Eddie Rubin once joked 'if we can generate 200kb reads you (EE) will finally shut up' #AGBT15
9:30am February 26th 2015 via Hootsuite
RT @jgreid: EE: many contigs shatter right over segmental duplications, justifies a de novo approach #AGBT15
9:29am February 26th 2015 via Hootsuite
EE: Summary - 'we are missing most of the variation from 50bp - 2kbp in size'. Estimate >50% can be genotyped if we can discover it #AGBT
EE: Shows a large table of ~18 ongoing projects, targeting 20x to 60x, across 4 institutions #AGBT15
9:28am February 26th 2015 via Hootsuite
EE:De novo human genome assy: recommend a hybrid approach with targeted BAC clone-based sequence #AGBT15
9:27am February 26th 2015 via Hootsuite
EE:W/a relatively few number of samples char in this way, can get a much better reference. Can use BACs to resolve complex regions #AGBT15
9:25am February 26th 2015 via Hootsuite
EE:3.9kb deletion from TGP data - now can go back and genotype, find 10% of TGP samples have this deletion #AGBT15
9:24am February 26th 2015 via Hootsuite
EE:Now at work on CHM13 w/P6C4 chemistry; increase in read length as distribution. Similar bar chart. Huddleston et al unpubl #AGBT15
9:22am February 26th 2015 via Hootsuite
EE:Retrotransposon biases - 20% inc in detection of SVA elements (repeat + GC composition) #AGBT15
9:19am February 26th 2015 via Hootsuite
EE:Relevance - implications of STRs in disease. Resolving complex repeats - very busy slide, illus. the complexity of the task #AGBT15
9:18am February 26th 2015 via Hootsuite
EE:Longer and new STRs - 3x more and 2.8x longer in CHM1 cp to reference genome #AGBT15
EE:Alu and L1HS sites. 22.1k novel var's correspond to 11Mbp of sequence. 6.7k events in 3.4k genes, 169 w/in coding regions #AGBT15
9:17am February 26th 2015 via Hootsuite
EE:Despite thousands of genomes sequenced, most of this var 50bp-5kbp 'never seen before' #AGBT15
9:16am February 26th 2015 via Hootsuite
EE:Fn relevalence - 20 exons, containing 12 genes. Payoff slide: event size against count. 92% ins, 60% dels are novel 50bp-5kbp #AGBT15
9:15am February 26th 2015 via Hootsuite
EE:'Signif. enrichment of tandem repeats'. Sometimes >10kb; flanked w/GC-rich on both sides. Sequence tech marched through it #AGBT15
9:14am February 26th 2015 via Hootsuite
EE:Summarizing: 90 gaps closed or extended; 26K 'structural var's' resolved at bp level resolution; 460 STR/VNTR >1kb #AGBT15
9:13am February 26th 2015 via Hootsuite
EE:An enucleated egg, 45.8x via RSII P5/C3 data released in late 2013 here: http://t.co/o6Qi3I5XeX Ref: http://t.co/x3HMrC3WdD #AGBT15
9:12am February 26th 2015 via Hootsuite
EE:Non-random 15% error is a feature, due to lack of bias. CHM1 hydatidiform mole is the 'Platinum Genome Assy' - not quite finished #AGBT15
9:10am February 26th 2015 via Hootsuite
EE: @PacBio histogram: 10.8kb mean, max 47.6kbp, very little G-C bias demonstrated "almost flat" cp to PCR-free illumina sequence #AGBT15
9:09am February 26th 2015 via Hootsuite
EE: 'We don't know how big that whole is'. Introduces @PacBio, they've been using it for 4y. ZMW's sequence directly (3rd gen) #AGBT15
9:08am February 26th 2015 via Hootsuite
EE: Missing heritability may be hiding in this missing variation due to short-read methods. 'We want to be comprehensive & precise' #AGB
9:07am February 26th 2015 via Hootsuite
EE: Structural var detection is indirect: read depth, read pair, split reads: all dep. on context Nature Rev http://t.co/Q4XdYaCs6w #AGBT15
9:06am February 26th 2015 via Hootsuite
EE: 1,000 genes in complex areas not well char. Alkan Nature Meth 2011 'we've been seduced' by NGS http://t.co/ZUC34dqd2v #AGBT15
9:05am February 26th 2015 via Hootsuite
EE: Most attn on SNV's & indels; today structural var >50bp, microsatellites, mobile elements 300bp-10kb #AGBT15
9:03am February 26th 2015 via Hootsuite
First: Evan Eichler, University of Washington: Resolving the Complexity of Human Genetic Variation by Single-Molecule Sequencing #AGBT15
9:02am February 26th 2015 via Hootsuite
The US Supreme Court has affirmed that fish are not an information-storage device - Quartz http://t.co/sbXZUX2NAv
7:15am February 26th 2015 via Hootsuite
Beautiful beach walking here at #AGBT15, although rain in the forecast. http://t.co/cZ98rXncVq
7:05am February 26th 2015 via Hootsuite
Star talker: Neil deGrasse Tyson on fame, education and tweets - The Washington Post http://t.co/N0xJceQd76
6:15am February 26th 2015 via Hootsuite
.@howitts_done Here's the NEJM ref ($) http://t.co/fVVdZVAoug 'myeloid cancers' (general).
5:26am February 26th 2015 via Hootsuite in reply to
Eating peanuts prevents peanut allergies http://t.co/3n4rH984gg
5:10am February 26th 2015 via Hootsuite
