Bob Daber (Gnosity Consults) Pushing the limits - challenges of somatic variant detection #Tricon

7:51pm February 23rd 2017 via Hootsuite

Mockus: Future, Maine Cancer Genomics will sequence 1800 pts over 3yrs, mech for clinical review https://t.co/0WSgAU4Be5 #Tricon

7:44pm February 23rd 2017 via Hootsuite

Mockus: Drug classes controlled vocabulary for target spec, hyperlinked to PubMed sources Glossary: https://t.co/22HIgszZTZ #Tricon

7:42pm February 23rd 2017 via Hootsuite

Mockus: Setup Cron jobs for automated updates. Long table of decision matrix posted here: https://t.co/Zn67kemymE #Tricon

7:39pm February 23rd 2017 via Hootsuite

Mockus: Even PDX and other functionalities (unmet needs, expression etc.) Easy to build, but hard to keep updated. #Tricon

7:37pm February 23rd 2017 via Hootsuite

Mockus: Genes, variants, clinical trials; Rx approach, disease oncology; CNVs, SNVs, Fusion, expression data is all there. #Tricon

7:36pm February 23rd 2017 via Hootsuite

Mockus: @jacksonlab has made their CKB open to the public, for not-for-profit use https://t.co/5ZZxofquU6 (Amazing. Kudos!) #Tricon

7:35pm February 23rd 2017 via Hootsuite

Mockus: JAX has a CKB (clinical KB, literature-based), also patient curated kb (tumor type and alteration data) #Tricon

7:34pm February 23rd 2017 via Hootsuite

Mockus: Guidelines also from AMP/CAP, '17 JMD https://t.co/XnDcN5zbfc #Tricon

7:31pm February 23rd 2017 via Hootsuite

Mockus: Guidelines came out recently: WashU's CiVIC '17 Nature Gen https://t.co/BfmWr1nZKY MSKCC too #Tricon

7:30pm February 23rd 2017 via Hootsuite

Mockus: 2nd need is guidelines ofr interpretation (actionability) - but what does that include? #Tricon

7:28pm February 23rd 2017 via Hootsuite

Mockus: First need: nomenclature - SNV, SNP, Ins, Del, indel, dup. It can be an accounting problem (+/- strands etc) #Tricon

7:27pm February 23rd 2017 via Hootsuite

Mockus: How to find which somatic var's important? 400 gene panel may have 2000 var's, 300 protein, 25 db mapped, 3-5 'actionable' #Tricon

7:26pm February 23rd 2017 via Hootsuite

Mockus:and Names of vendors with connections: QIAGEN, N-of-One, ArcherDx, PierianDx etc. about 20-25. #Tricon

7:24pm February 23rd 2017 via Hootsuite

Mockus: Showed seven major nodes as a network, not very well connected. Vendors: bioinf, content, dx, integrator, platforms #Tricon

7:23pm February 23rd 2017 via Hootsuite

Mockus: Did a data map of data silos, color coded by type (pharma, cancer, non-profit, dx lab etc) and size - interesting! #Tricon

7:23pm February 23rd 2017 via Hootsuite

Susan Mockus (Jackson Laboratories CT) Creating standards and transparency for interpretation of somatic variants #Tricon

7:22pm February 23rd 2017 via Hootsuite

Zimmerman: Next: RUO, optimize process, costs; std protocol to remove re-pooling for each bespoke panel. H2 '17 launch internal #Tricon

6:30pm February 23rd 2017 via Hootsuite

Zimmerman: May be able to ID adjuvant chemotherapy-sensitive disease. cfDNA branch rise may predict metastatic recurr #Tricon

6:28pm February 23rd 2017 via Hootsuite

Zimmerman: Showed SCC StgI pts, 15/16 able to pick up trunk mutations; 50% detected on the trees. Levels clearly lower. #Tricon

6:25pm February 23rd 2017 via Hootsuite

Zimmerman: Showed large sens/spec chart. Onto TRACERx CRUK to monitor lung-cancer evolution, building trees https://t.co/bjYzKFKmVN #Tricon

6:24pm February 23rd 2017 via Hootsuite

Zimmerman: Per-sample cost is on the order of $100; showed assay sensitivity by spike titrations; 99% sens at 0.1% AF #Tricon

6:20pm February 23rd 2017 via Hootsuite

Zimmerman: Targeted 20-30 SNVsper pt (>100 possible) for max sens. 40K read depth; use synth mutant spikes, 33ng input #Tricon

6:17pm February 23rd 2017 via Hootsuite

Zimmerman: SNV sens claimed single-cell down to 20% AF #Tricon

6:15pm February 23rd 2017 via Hootsuite

Zimmerman: Shows T21 data from gDNA and single-cell; 28k-plex mmPCR; shows Br ca bulk and single-cell very similar #Tricon

6:15pm February 23rd 2017 via Hootsuite

Zimmerman: Natera's Panorama 'fastest growing molecular test in history' Detect CNV to 2.5% https://t.co/kH6igjQKfD #Tricon

6:13pm February 23rd 2017 via Hootsuite

Zimmerman: Peformed over 300K tests last year; CLIA-lab in San Carlos CA; launched their first test in '11 (paternity), '13 for NIPT #Tricon

6:11pm February 23rd 2017 via Hootsuite

Bernard Zimmerman (Natera CA) Personalized custom panels for the detection of circulatory tumor DNA mutations #Tricon

6:10pm February 23rd 2017 via Hootsuite

Mason: Federated query structure: single db shared across the UK #Tricon

5:55pm February 23rd 2017 via Hootsuite

Mason: For interpretation, lists WuxiNextCODE, Illumina, a few others. Has a BCIP (Bioinf. Clin Interpret. Partnership) #Tricon

5:54pm February 23rd 2017 via Hootsuite

Mason: Genes: 186 gene panels, 3709 genes, and browse-able here: https://t.co/5Adyl5N1Aw #Tricon

5:53pm February 23rd 2017 via Hootsuite

Mason: 'Databases are not very good when the knowledge is in people's heads' Thus they crowdsource expertise #Tricon

5:51pm February 23rd 2017 via Hootsuite

Mason: 200 disorders are nominated as std of care; eligibility criteria established; data model std'ized for clinicians #Tricon

5:50pm February 23rd 2017 via Hootsuite

Mason: FF designated as an official 'diagnostic source' as a consensus 'Human Tissue Authority' statement (!) #Tricon

5:47pm February 23rd 2017 via Hootsuite

Mason: Sequencing cancer genomes: decided that FF tissue instead of FFPE; embedded the processing of FF by the pathologists #Tricon

5:46pm February 23rd 2017 via Hootsuite

Mason: 'delivery partners' include Seq Centr in Hinxton and MRC to handle biorepository; GE Clinical Interpret. Partnership (GeCIP) #Tricon

5:39pm February 23rd 2017 via Hootsuite

Mason: Driving toward a coordinated response with a unified healthcare system. 13 GMCs (Genomic Medicine Ctrs) w/3-5M people #Tricon

5:37pm February 23rd 2017 via Hootsuite

Mason: To leave a legacy of a database that is useful to the NHS. 50% cancer, tumor/germline. 21PB of data (2K years of mp3's) #Tricon

5:35pm February 23rd 2017 via Hootsuite

Mason:Also ethically transparent programme based on consent; kick-start the dev of a UK genomics industry, trials, new Dx #Tricon

5:33pm February 23rd 2017 via Hootsuite

Mason:Objectives are: rare inh disease, cancer, pathogens; referring new diagnoses to NHS (new medical insight) #Tricon

5:33pm February 23rd 2017 via Hootsuite

Mason: David Cameron in Dec '13. Setup July '13, working group Prof Sally Davies est a strategic priorities working gp #Tricon

5:32pm February 23rd 2017 via Hootsuite

Joanne Mason (Genomics UK) The 100,00 Genomes Project: transforming the UK's National Health Service #Tricon

5:30pm February 23rd 2017 via Hootsuite

Lek: Re-assessment of previous reports of pathogenicity: abnormal liver fn. But var >1% in latino population, used to reassess #Tricon

5:15pm February 23rd 2017 via Hootsuite

Lek: Graphed diversity of their dataset; able to go from 11.9k down to 132 rare missense or truncating. Found likely causative gene #Tricon

5:11pm February 23rd 2017 via Hootsuite

Lek: Graphing haploinsufficiency, figure from '16 Nature https://t.co/HpxPCXKvpt #Tricon

5:09pm February 23rd 2017 via Hootsuite

Lek: ID genes with significant depletion of varation: observed vs expected; loss of function dpressed observations #Tricon

5:08pm February 23rd 2017 via Hootsuite

Lek: Scalable pipeline for joint discovery: from 925TB of BAMs to 3.3TB of final VCFs. CPU days: 15.6K days, 950K, then 13.6 at VCF #Tricon

5:05pm February 23rd 2017 via Hootsuite

Lek: ExAC aggregated and called 92K exomes; with a subset of 60,706 'reference' samples (high-quality, unrelated, consented) #Tricon

5:03pm February 23rd 2017 via Hootsuite

Lek: When he arrived a few years ago: 500K exomes have been sequenced; silo'ed by project, inconsistently processed. #Tricon

5:02pm February 23rd 2017 via Hootsuite