Murphy: (Showed a troubleshooting exercise with inconsistent promoter CpG island signal) Believes there's bias in their WGA #IWT14
5:12am September 12th 2014 via Hootsuite
Murphy: Replicates have R about 0.6; showed cross-method comparison; compared to HiSeq platform (very similar) #IWT14
5:06am September 12th 2014 via Hootsuite
Murphy: Showed data, has done 24 samples across 24 PI chips; shows good vs poor read distribution (i.e. bimodal short ones) #IWT14
Murphy: Maps to mouse reference, filter reads <50bp, normalization, 200bp non-overlapping read counts across genome #IWT14
5:05am September 12th 2014 via Hootsuite
Murphy: His post-doc 'data within a week' 'he's thrilled with this' due to speed of Proton #IWT14
5:04am September 12th 2014 via Hootsuite
Murphy: Mouse DNA, uses a modified meDIP protocol; single strand to WGS after pull-down; modified library prep #IWT14
5:03am September 12th 2014 via Hootsuite
Murphy: Reduction of 5hmC importance in cancer; modification of protocol (Thomson 2013 NAR). 3 methods compared #IWT14
5:02am September 12th 2014 via Hootsuite
Murphy: Points out methylation in the gene body; Peter Laird's work (Epigenetics rev 2013) #IWT14 BtB Blog http://t.co/y42NVx8fGC
5:01am September 12th 2014 via Hootsuite
Murphy: #IWT14 Refers to the importance of 5hmC, an overlooked methylation mark http://t.co/nO07kQpuDz
5:00am September 12th 2014 via Hootsuite
Murphy: Methylation via sequencing can distinguish between 5mC and 5hmC; other adv in chart (comparing microarray to NGS) #IWT14
4:57am September 12th 2014 via Hootsuite
Murphy: Many 450K methylation arrays; recent work on Crohns (PubMed) http://t.co/5jTQiHE5xn #IWT14
4:56am September 12th 2014 via Hootsuite
Murphy: "I really like the flexibility" of the Ion Torrent platforms. Ion Chef "It's great" #IWT14
4:54am September 12th 2014 via Hootsuite
Murphy: They have biobanking, sample processing, and analysis (historically genotyping via qPCR, microarrays, more on PGM / Proton #IWT14
Lee Murphy (Wellcome Trust, Western Hosp. Edinburgh) "ChIP-Seq analysis on the Ion Proton System" #IWT14
4:52am September 12th 2014 via Hootsuite
Mangion:Q: Change in reference db's? A: You can upload references in TS; in Ion Reporter does take time, when it stabilizes #IWT14
4:48am September 12th 2014 via Hootsuite
Mangion: Has 'Training Beyond the Box' - Genomics in Medicine, Bioinf. for molecular biologiests #IWT14 http://t.co/NsxWiIB6R9
4:46am September 12th 2014 via Hootsuite
Mangion: Points out availability of IT Professional Services; obstacles to implementation disappear when they get involved #IWT14
4:43am September 12th 2014 via Hootsuite
Mangion: A balance of sensitivity (low stringency) and specificity (high stringency); one size does not fit all #IWT14
4:39am September 12th 2014 via Hootsuite
Mangion: ("dept" in prior post = "dependent") #IWT14
4:38am September 12th 2014 via Hootsuite
Mangion: Mentions NIST Genome in a Bottle; showed sens shooting up dept on var caller #IWT14 Blog post: http://t.co/jTjOTYbbCD
4:37am September 12th 2014 via Hootsuite
Mangion: Novel insertion - example filtered out due to insufficient coverage (11x instead of default param 15x) #IWT14
4:36am September 12th 2014 via Hootsuite
Mangion: Sometimes annotations in db's as 'hotspot'; a putative del in homopol on L end. Filtered out due to 'common signal shift' #IWT14
4:34am September 12th 2014 via Hootsuite
Mangion: Mentions HiQ: for smaller panels, software is already dealing with indels like these, so improvement on these panels limited #IWT14
4:32am September 12th 2014 via Hootsuite
Mangion: Strand bias (137 alt allele of 1200) + position of del in homopol = evaluated and rejected #IWT14
4:31am September 12th 2014 via Hootsuite
Mangion: "Source:novel" is a DEL, coverage ~1200, shows strand bias, more info in flow space for more detail. Del in homopol 5x "G" #IWT14
4:30am September 12th 2014 via Hootsuite
Mangion: Another "Source:Hotspot" forces a genotype call, but no data is there (homozygous to ref), 0% freq #IWT14
4:29am September 12th 2014 via Hootsuite
Mangion: "Source: Hotspot" goes straight to evaluation; example GG to TT at 2.3%, quality score is relatively low #IWT14
4:28am September 12th 2014 via Hootsuite
Mangion: 'Source: Novel' variant is a lablel when it lies outside a hotspot region #IWT14
4:27am September 12th 2014 via Hootsuite
Mangion: "What is the hypothesis to test?" is the consideration - what's involved with changing parameters in TVC #IWT14
4:24am September 12th 2014 via Hootsuite
Mangion: TVC (variant caller) Evaluation of candidate var's have many parameters; many caveats to change from defaults #IWT14
4:23am September 12th 2014 via Hootsuite
Mangion: Torrent Variant Caller has block polymorphism feature, although turned off by default (computationally intensive) #IWT14
4:20am September 12th 2014 via Hootsuite
John Mangion (Thermo Fisher Scientific) "Variant calling, mutation analysis and data integration" #IWT14
4:18am September 12th 2014 via Hootsuite
Sharkey: Ex.: for Colon/Lung v2, >700 samples worth of data compiled among the collaborators #IWT14
4:15am September 12th 2014 via Hootsuite
Sharkey: Implementation networks are three or more labs that he coordinates; a flexible, adaptable approach, 8-12w timeframe #IWT14
4:14am September 12th 2014 via Hootsuite
Sharkey: List of early FALCON in development: RASopathies, MDS, others #IWT14
4:13am September 12th 2014 via Hootsuite
Sharkey: A single AmpliSeq Colon/Lung run (16 barcoded samples on a 318 chip) is 1472 CE reactions #IWT14
4:09am September 12th 2014 via Hootsuite
Sharkey: How to remove the barriers? A jigsaw puzzle of cost, workflow, throughput, performance, bioinformatics. #IWT14
4:07am September 12th 2014 via Hootsuite
Sharkey: Helpful reminder that NGS is only <10 years old. Listed many of the applications on CE from MLPA, fragment analysis, etc. #IWT14
Freddie Sharkey (Thermo Fisher Scientific) "CE to NGS on the PGM and the role of implementation networks" #IWT14
4:06am September 12th 2014 via Hootsuite
Wood: PGM Torrent Server 12TB, Proton TS 24TB. http://t.co/FNfOkmEe8C gives 100GB data free #IWT14
4:01am September 12th 2014 via Hootsuite
Wood: Data structure is DAT, WELLS, BAM, Aligned BAM, VCF: files to backup are BAM and VCF's #IWT14
4:00am September 12th 2014 via Hootsuite
Wood: Amplicon coverage visuallized with a click - on any gene of interest across samples #IWT14
3:54am September 12th 2014 via Hootsuite
Wood: Filter chain automation in Ion Reporter: from 54K var's (exome sample) to 1 with 3 filters shown #IWT14
3:52am September 12th 2014 via Hootsuite
Wood: Variant lists can be saved - to either quickly filter in (a quick check) or filter out (common polymorphisms, possible FPs) #IWT14
3:50am September 12th 2014 via Hootsuite
Wood: Day to day, setup of workplan template (with barcodes), Oxford found that Ion's param. optimizations were already 'very good' #IWT14
3:49am September 12th 2014 via Hootsuite
Up next: Joe Wood (Thermo Fisher) "Bioinformatics for Clinical Research" #IWT14
3:43am September 12th 2014 via Hootsuite
With genetic testing, I gave my parents the gift of divorce - Vox http://t.co/ZsQ1ArIXlN HT @notsojunkdna 'a must-read' about @23andMe
10:00pm September 11th 2014 via Hootsuite
Apple's Steve Jobs was concerned about his children's gadget use - Telegraph http://t.co/hBNmdTRdtX
9:01pm September 11th 2014 via Hootsuite
The Internet Probably Isn't Ruining Your Teenager's Brain | WIRED http://t.co/I1wuiGHLc5
8:01pm September 11th 2014 via Hootsuite
The newest sequencer on the block: Base4 goes public | CoreGenomics http://t.co/qz9b0fwnU9
7:00pm September 11th 2014 via Hootsuite
