#ampeurope2018 Hoffman: Clinical trials progressing - but slowly. Data fed into supervised, unsupervised, and deep learning models
9:23am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Hoffman: 23M citations in Medline, and major opps to improve quality of care, HAI, AEs, disparities
9:22am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Hoffman: We algorithms b/c there are limits to human memory; limits to perception (170 pixels/inch),… https://t.co/vwXMaEvBUW
#ampeurope2018 Hoffman: Many goals and has elusive goals - predict and improve pt outcomes; improve utilty; demo cost-effectivenss
9:21am May 2nd 2018 via Twitter Web Client
Mark Hoffman (Children's Mercy Kansas City MO, USA) #ampeurope2018 Turning Little Data into Big Data: Data Science and the Omics
9:17am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Laura Tafe (Dartmouth NH) Emerging knowledge: clinically significant big data
#ampeurope2018 Ferreira-Gonzalez: You need trained individuals to interpret in the clinical context. And policies for consistency needed.
7:02am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Dinjens: TERT promoter has high GC, they used a hotspot file for Battle4 sample finding. Value of te… https://t.co/hvxyXWHzEm
7:01am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Winans Dinjens (Erasumus Med Ctr Rotterdam NE): their own group doesn't know. Shows summary result b… https://t.co/KjWs9tQfIL
6:59am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Tyrelle: Battle5: 26 clinical trials, many pages of reporting (and available as demo at their stand)
6:53am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Tyrelle: Reviews Battle1 data: nothing special in the output; shows report for Battle2/3 data; Battl… https://t.co/auv4RdglH6
6:52am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Tyrelle: Illustrates flow info and how its used: where FNs get special information. Flow-space numbers: fwd/rev strands
6:46am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Tyrelle: Shows very long list (30?) of parameter settings 'sensible defaults'. BRCA-spec: has hotspo… https://t.co/Eap4NNYUgi
6:44am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Tyrelle: Torrent Variant Caller (TVC) have two flows: one w/o hotspots, another with. W/o: no calls may be there.
6:43am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Tyrelle: And Oncomine KB Reporter links variants to evidence. Reporter is free to use to a certain amt.
6:40am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Tyrelle: And larger chips have exome capability. Aligned BAM gets to variant calls. Annotate and Filter: using Ion reporter
6:39am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Tyrelle: Imptc of familiarity. Having BAM files is important, as well as FASTQ. Shows Ion method. Io… https://t.co/B2oFyfTOHM
#ampeurope2018 Greg Tyrelle @gtyrelle Thermo Fisher: other competitors (5 others) declined to participate.
6:38am May 2nd 2018 via Twitter Web Client
Q: Take into acct strand bias? #ampeurope2018 Bonnert: Many automatically removed but for one mutation in question a manual component
6:29am May 2nd 2018 via Twitter Web Client
Q: Does their literature citations get weighted? #ampeurope2018 Bonnert: Yes they do weight and other factors (drug approval)
6:27am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Bonnert: Standards for minimum thresholds, and what rules to establish. Working w/data impt, assessing real/not real
6:24am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Bonnert: Battle4 TERT in context of NSCLC pathogenic in this context. Battle5: have TP53 that is called germline
6:22am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Bonnert: Battle3 is Tier1A Pathogenic, EGFR from NSCLC. Battle4 is Tier 2C. Pathogenicity is context… https://t.co/S5avMOjMiE
6:18am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Bonnert: And their local re-alignment of reads, not seen in population databases.
6:16am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Bonnert: Actionability, pathogenicity. Goes through each sample. Battle1: Tier 2C. Battle2: CDKN2A del was not in aligned BAM
#ampeurope2018 Bonnert: There is a need for standards: like min depth, target region vs reportable regions.
6:12am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Bonnert: Imported FASTQ, align, call, filter <5% VAF. After QCI: added sample Dx, tumor cell %
6:10am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Bonnert: HGMD, and Ingenuity Pathway Analysis. QCI works on ACGM/AMP, AMP/ASCO/CAP guidelines/Rx
6:09am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Bonnert: Have a 'biomed genomics workbench' multi-tech secondary analysis; and interpretation includ… https://t.co/RQlSeA65U5
6:08am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Bonnert: The goal is a clear report. Reality: is complex pipeline scripts, datasources, dataformats.
6:07am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Tim Bonnert QIAGEN - not a clinical geneticist as a disclaimer. QCI interpret 'for illustration and educational use'
6:06am May 2nd 2018 via Twitter Web Client
Hage #ampeurope2018 If known data comes back; if unknown will be manually curated and come back in. A partnership with N-of-One
6:04am May 2nd 2018 via Twitter Web Client
Q: Report doesn't show FDA trials? Import back in? #ampeurope2018 Hage: that is the plan, to get an xml file back to add in.
6:03am May 2nd 2018 via Twitter Web Client
#ampeurope2018 If you use Agilent panels, aligner/caller pushes directly in. "Is it free to use? (no answer)"
6:02am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Marked CE-IVD, Class I exempt in US, Class I in AUS
5:56am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Shows summary table of all 5 Battle samples; sample 4 has potential signficance. N-of-One partnership: manually curated vars
#ampeurope2018 'all-in' approach to calling, or 'consensus calls between aligners' more restrictive.
5:54am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Can check for fn predictions. N-Of-One report connected, finer level of detail. Align and call prefe… https://t.co/QaZpd0d0PT
#ampeurope2018 COSMIC, CIViC, ClinVar, their own MVL database. Check for coding effect (stop loss/gain), or splice site mut
5:49am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Called 221 vars, narrowed to 18; then took 500x readdepth and quality threhshold. Can change reference.
5:47am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Shows the default report. Shows complex filtering tree 'we've designed to be easy to use'. Shows sel… https://t.co/ddjK0hfjVd
5:46am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Any kind of VCF, platform agnostic. Can use various callers. Interprets SNV, CNV, indels, SVs.
5:44am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Decision support across applications (from inh, carrier screen, germline cancer, somatic cancer). Across assay types
#ampeurope2018 Agilent to start: Elias Hage: Alissa Interpret is the newest Cartagenia Workbench. A reporting tool.
5:43am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Battle1/5 labeled datasets: .bed, 5 FASTQ, 5 BAMs. Sr tech Ron van Marion, bioinformatics Niels Krol.
5:41am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Asked for var's, annotation, AF, interpret of vars/pathogenicity on 5 tiers, and whether clonal or subclonal
5:38am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Aimed at clonal/subclonal mutations (not on low level <5% VAF), data generated by Ion Torrent 5XL platform
#ampeurope2018 5 indep tumor samples, routine FFPE 'nothing special', analyzed conventionally. Tumor type, and tumor cell %
5:37am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Elias Hage Agilent, Tim Bonnert QIAGEN, Greg Tyrelle @gtyrelle from Thermo Fisher are the participants.
5:36am May 2nd 2018 via Twitter Web Client
#ampeurope2018 Battle of the Bioinformatics Pipeline: background is managing, filtering, reporting data. Asked companies to help
5:35am May 2nd 2018 via Twitter Web Client
