A4 Druley: Multiplexed ddPCR isn't cheap nor easy; orthogonal platform with same LOD #AMP2017
9:45am November 17th 2017 via Hootsuite
A3 Druley: Tries to start with 500K template molecules / sample. 1:10K you need to have enough molecules to start with #AMP2017
A2 Druley: 8-oxo-guanine due to oxidative damage in extraction, G to T. '15 paper was 1:300 to 1:500 for G to T change #AMP2017
9:44am November 17th 2017 via Hootsuite
Q: For MRD, how consistent over time? Validation for new var's? Druley: The answer is no; expected C to T b/c of deamination #AMP2017
9:43am November 17th 2017 via Hootsuite
A2: Druley: Also rules around read families to ensure fidelity of UMI identifier #AMP2017
9:42am November 17th 2017 via Hootsuite
Q: Errors in the UMI? Druley: Looking at redundancy - errors in the index. Only need 3-7 reads/index. An early 'new… https://t.co/Fy19o4uCij
Druley: Ped AML - 'we've made progress but at a cost - delayed learning and growth'. Credits Illumina, Archer. Inde… https://t.co/4NAeOXrFjk
9:40am November 17th 2017 via Hootsuite
Druley: Goal - build comprehensive 'sequencing toolbox to improve MRD detection and precision drug selection for ped AML' #AMP2017
9:39am November 17th 2017 via Hootsuite
Druley: custom Archer 40 genes, 28kb, also Archer Heme v2. Also w/CHOP - proteomics. Cp against MRD via flow #AMP2017
9:38am November 17th 2017 via Hootsuite
Druley: Earlier finished largest AAML 1531, 870 dx, 870 EOI1, 170 relapse 1,900 samples. 80 genes, 200kb custom ILM myeloid. #AMP2017
9:37am November 17th 2017 via Hootsuite
Druley: DNA, RNA, imaging, cyto, age, gender... use machine learning to use PCA and stratify #AMP2017
9:36am November 17th 2017 via Hootsuite
Druley: Fish plots: tracking clonal expansion. Circos: if you can squint... you'll 'go nutty if you look at them too much' #AMP2017
9:35am November 17th 2017 via Hootsuite
Druley: Shows interesting visual data rep - reactome, fish plot, circos. Ph+-like ALL: how can this be best shown? #AMP2017
Druley: TARGET initiative ('TCGA for ped ca'): EOI2 AML samples, only 1 known fusion via 'COG'. Able to detect fusi… https://t.co/3l0XoWTnEd
9:32am November 17th 2017 via Hootsuite
Druley: Unk whether RUNX1 abnormal transcript is clin significant, but interesting biology that couldn't be detected previously. #AMP2017
9:31am November 17th 2017 via Hootsuite
Druley: Looking at structural diff in RNA - back to ArcherDx HemeV2, looked at splice muts in RUNX1. #AMP2017
9:30am November 17th 2017 via Hootsuite
Druley: Have applied UMIs to RNA, collab with QIAGEN. All genes <100 count transcripts. Able to get counts to 3.5 c… https://t.co/zoZ5nHH
9:29am November 17th 2017 via Hootsuite
Druley: 47 clonal types pre-Rx. But post-chemo they disappear. Post-chemo 30 new ones appear #AMP2017
9:28am November 17th 2017 via Hootsuite
Druley: Using de novo assembly, 'LOD 0.001' or 0.1%. #AMP2017
9:25am November 17th 2017 via Hootsuite
Druley: Looking at FLT3-ITD, working with ArcherDx explains single-primer method for multiple different size amplicons #AMP2017
9:24am November 17th 2017 via Hootsuite
Druley: From close analysis: found four mutations (DNMT3A 2x), ASXL1, JAK2 with p-values <0.01, up to 15y in advanc… https://t.co/kF5oV8X
9:22am November 17th 2017 via Hootsuite
Druley: no difference in overall mutation burden between healthy and those who develop disease. Showed plot of DNMT3A muts #AMP2017
9:19am November 17th 2017 via Hootsuite
Druley: Were these just quiescent cells? 40y old blood examined via flow: data suggest normal hematopoetic aging #AMP2017
9:17am November 17th 2017 via Hootsuite
Druley: Showing rise of mutations over time across 5 samples, 10y: appear to be self-renewing (!). 2nd and 3rd hits… https://t.co/4krJhA4UcI
9:16am November 17th 2017 via Hootsuite
Druley: Demonstrated 95% of samples had clonal mutations, in contrast to prior work. 40% in DNMT3A and TET2. #AMP2017
9:15am November 17th 2017 via Hootsuite
Druley: Using 20 samples from healthy elderly nurses (Nurses' Health cohort) https://t.co/megBpCj4GK validated to 0.1% #AMP2017
9:14am November 17th 2017 via Hootsuite
Druley: In order to expand (up to date only known mutations), used TruSight Myeloid panel, and added UMI's #AMP2017
9:12am November 17th 2017 via Hootsuite
Druley: Published their error-corrected method in '15 https://t.co/dzhVj02BOH #AMP2017
9:11am November 17th 2017 via Hootsuite
Druley: 20y of samples of AML pts at WashU; track TP53 mutations as pre-existing and selected for via chemo #AMP2017
9:10am November 17th 2017 via Hootsuite
Druley:GATA1 for ped cancer in Trisomy 21 is common, showed 5bp ins downt to 1:10,000 #AMP2017
9:09am November 17th 2017 via Hootsuite
Druley: Points out limitations of regular NGS (0.5% to 2% limit of detection), and have developed a 16bp UMI method #AMP2017
Good morning! Todd Druley (WashU) Single Molecule Quantification of Rare DNA and RNA Variants in Heterogeneous Samples #AMP2017
9:08am November 17th 2017 via Hootsuite
Koboldt: 'Database-driven methods (eg COSMIC for TPs and dbSNP for FPs) are increasingly unreliable' #AMP2017
3:55pm November 16th 2017 via Hootsuite
Koboldt: Ideally: precise, few FPs: establish 'ground truth' w/orthogonal valdatoin, manual review, or in silico datasets #AMP2017
Koboldt: Shows signal-to-noise ratio - 'not entirely random, not entirely uniform'. Accurate est of VAF requires deep NGS coverage. #AMP2017
3:53pm November 16th 2017 via Hootsuite
Koboldt: Shows VarScan's calculation of Fisher's Exact Test. Solid tumor challenges: tumor purity; for liquid tumor… https://t.co/03o3a4o5fC
3:52pm November 16th 2017 via Hootsuite
Koboldt: Makes point that tumor/normal pairs should be used, under same protocol, similar conditions. #AMP2017
3:47pm November 16th 2017 via Hootsuite
Koboldt: And transcriptome: may not be the best for somatic discovery. Powerful to use as complement #AMP2017
Koboldt: NGS strategies - WGS; WES/panel capture; Transcriptome contrasted. Comprehensive but costly; senstitive/ef… https://t.co/zgwfAfnjND
3:45pm November 16th 2017 via Hootsuite
Daniel Koboldt (Nationwide Children's Hosp Columbus OH) Strategies and challenges for somatic mutation detection by NGS #AMP2017
3:44pm November 16th 2017 via Hootsuite
Mullighan: Looking at secondary lesions: IKZF1 '15 ref https://t.co/rIWTcT1JZs find de-repression of cellular programs: adhesion #AMP2017
12:40pm November 16th 2017 via Hootsuite
Mullighan: Data suggests 'transformation of early progenitor' driving phenotype. #AMP2017
12:38pm November 16th 2017 via Hootsuite
Mullighan: Shows gene expression from ZNF384-rearr in B/myeloid mixed. Lymphoid and myeloid lineage splits. #AMP2017
Mullighan: New subtypes - MEF2D-rearr B-ALL, RNA-seq of 560 B-ALL cases, multiple recurring targets MEF2D, ZNF384, NUTM1, IGH. #AMP2017
12:34pm November 16th 2017 via Hootsuite
Mullighan: #AMP2017 '17 review of Ph+like ALL https://t.co/NMqUwbGwKe
12:33pm November 16th 2017 via Hootsuite
Mullighan: Shows mouse data showing feasibility for TKI with std therapy. COG AALL1131, COG AALL1521 clinical trials. #AMP2017
12:32pm November 16th 2017 via Hootsuite
Mullighan: Druggability work: Over-expression and truncation of EPOR - several tyr residues, thus negative regulation of receptor #AMP2017
12:31pm November 16th 2017 via Hootsuite
Mullighan: 'As we sequence more we find more kinases and rearrangements'.... '12 ref https://t.co/e2OltwkhTG #AMP2017
12:30pm November 16th 2017 via Hootsuite
Mullighan: Classes - JAK-STAT, ABL, other signaling including STAT MAPK. Pathways raise possibility of multiple drugging #AMP2017
12:28pm November 16th 2017 via Hootsuite
Mullighan: Ph-Like ALL: similar transcriptome to Ph+ ALL. Then looked at prevalence. '17 JCO https://t.co/H1m6HN7kFW Most diverse #AMP2017
12:27pm November 16th 2017 via Hootsuite
