Homer: Where do you want to be on the ROC curve? Choice between sens and spec based on usage. #ionworld2013
9:40am October 22nd 2013 via Hootsuite
Homer: 'Truth' datasets, see the same variants are called different. NIST set has a well-char region with pos and negs #ionworld2013
9:39am October 22nd 2013 via Hootsuite
Homer: He writes his own tools (he is the author of TMAP); learnings about variant calling is 'variant zygosity matters'. #ionworld2013
9:38am October 22nd 2013 via Hootsuite
Homer: Same Claritas chart that Patrice showed yesterday - 5/8 'green', 3/8 'yellow', including Speci. Indels, Spec.SNPs #ionworld2013
9:37am October 22nd 2013 via Hootsuite
Homer: Substitution error shines on the Proton, and total sensitivity is also very good Proton vs. Hiseq. #ionworld2013
9:36am October 22nd 2013 via Hootsuite
Homer: 50K variants/exome via AmpliSeq exome. Challenge areas - FP/Mb measurement (wish I took a photo!) #ionworld2013
9:35am October 22nd 2013 via Hootsuite
Homer: Caveat given about capture technologies; 90% target within 20% of the mean is 'very good'. 97% (1) and 92% (2) at >20x #ionworld20
9:33am October 22nd 2013 via Hootsuite
Homer: 70M good quality reads, 140bp mean RL, 135bp mean quality RL after trimming. 160x 1sample/run, 90x at 2sample/run #ionworld2013
9:32am October 22nd 2013 via Hootsuite
Homer: 10G throughput, >8.5G of AQ17 aligned, saw improvement in the fall. #ionworld2013
Homer: Six HapMap sample, aim for 80x min coverage, compared cross-platform between Proton and HiSeq #ionworld2013
9:31am October 22nd 2013 via Hootsuite
Homer: Uses NA12878 as a well-char reference, shout-out to Genome in a Bottle NIST effort. #ionworld2013
9:30am October 22nd 2013 via Hootsuite
Nils Homer, Claritas Genomics. 'Assessment of the Ion Proton System as tech platform for future clinical exome sequencing'. #ionworld2013
Thiede: Used cell-free DNA from plasma using fusion primer approach on PGM #ionworld2013
9:24am October 22nd 2013 via Hootsuite
Thiede: Using a fusion primer, increase input to 250ng template on PCR (40K haploid copies), dec # of PCR cycles. 250K reads #ionworld2013
9:20am October 22nd 2013 via Hootsuite
Thiede: MRD - tools to detect include qPCR, PCR, FACS, FISH. Using NGS for MRD detection; SNV detection at high sens. #ionworld2013
9:17am October 22nd 2013 via Hootsuite
Thiede: Minimal residual disease: 10x10^11 leukaemic cells of burden per patient. Know that majority will relapse. #ionworld2013
9:15am October 22nd 2013 via Hootsuite
Thiede: >95% coverage uniformity, >85% bases on-target, expected variants in 22 tested samples. In process of redesign #ionworld2013
9:14am October 22nd 2013 via Hootsuite
Thiede: Uses Ion Reporter for annotation of AML genes, 'quite useful' #ionworld2013
9:13am October 22nd 2013 via Hootsuite
Nice pics! RT @evennemeyer: Jun Wang and Milan Radovich wrapping up talks- day started with Eric Topol #ionworld2013 http://t.co/MrzY1Ixvcz
9:10am October 22nd 2013 via Hootsuite
Thiede: Group met to agree on 19 genes, well-characterized for AML. 2 pools, 337 primer pairs. Final list ~30 genes #ionworld2013
9:09am October 22nd 2013 via Hootsuite
Thiede: Now able to subdivide mutations to prognosis, and to apply it. Now a larger scale is needed - a research panel needed #ionworld2013
9:08am October 22nd 2013 via Hootsuite
Thiede: 'AML is a molecularly hetergeneous disease', Dohner et al Blood 2010 figure. #ionworld2013 http://t.co/VGaCnnnTvU
9:07am October 22nd 2013 via Hootsuite
Thiede: AML background - older (median 60y.o.) Dx by morphology, immunophenotyping, cytogenetic. Treatment mainly marrow xplnt #ionworld2013
9:04am October 22nd 2013 via Hootsuite
Up next: Dr. Christian Thiede, Dresden Univ of Tech, approaches for high-sensitive detection of molecular alterations in AML #ionworld2013
9:02am October 22nd 2013 via Hootsuite
Ray: "For 25 years I've worked with ABi, and I've always enjoyed working with them and it has always been very successful" #ionworld2013
9:00am October 22nd 2013 via Hootsuite
Ray: Want a push-button software solution, shows Ion Reporter screenshot and says it does the job. #ionworld2013
8:58am October 22nd 2013 via Hootsuite
Ray: Looking at a deletion in CF - obvious from reads #ionworld2013 http://t.co/EoM8YL78b9
Ray: 150 samples via CF consortia: 343/343 SNV all correct, 49/55 IVS8, 80/82 indels, sens. spec. both >99% #ionworld2013
8:55am October 22nd 2013 via Hootsuite
Ray: 74 samples, 6 ave var's/sample, 91% true pos, 6% false pos., 3% false neg. Will be tuning the software with LifeTech #ionworld2013
8:54am October 22nd 2013 via Hootsuite
Ray: 1k to 5k target coverage. >99% target at >30x, 'is excellent'. Exon 10 - homologous sequence, alignment discarded #ionworld2013
8:52am October 22nd 2013 via Hootsuite
Ray: "This test has to be cheap." "In a Dx setting it has to be fast." "There's no point... when you need 200 samples." scale #ionworld2013
8:50am October 22nd 2013 via Hootsuite
Ray: Now look for rare mutations, indels, intronic seq via PGM of all samples up-front (rather than a Luminex screen first) #ionworld2013
8:48am October 22nd 2013 via Hootsuite
Ray: CF testing currently on Luminex CF39, if 1 or 2 muts found will use MLPA dosage + whole gene sequencing (Sanger) #ionworld2013
8:47am October 22nd 2013 via Hootsuite
AmpliSeq #Ionworld2013 track: Peter Ray, Univ Toronto / Hospital for Sick Children. Comparison of CFTR between Sanger and Ion Torrent PGM
8:44am October 22nd 2013 via Hootsuite
RT @BioMickWatson: Got to love the rumour mill. That's the 3rd person mentioning the possibility of a PacBio benchtop instrument.
8:25am October 22nd 2013 via Hootsuite
For rare diseases RT @acmedsci: Our Fellows comment on the initiative to sequence 10,000 #genomes by @Cambridge_Uni http://t.co/e4ddDEBZRc
7:25am October 22nd 2013 via Hootsuite
IBM's Watson is better at diagnosing cancer than human doctors | Wired UK http://t.co/YiX8bpGCcT
6:49am October 22nd 2013 via Hootsuite
Gardner's top IT technology trends for 2014 | Forbes http://t.co/vXhxHfyqW4
5:25am October 22nd 2013 via Hootsuite
What can we expect for the next decade in Marketing. | Forbes http://t.co/5NkjU1pKoC
4:25am October 22nd 2013 via Hootsuite
RT @EricTopol: "The 0 gravity economics of #cancer" meds http://t.co/LcSibx5dYM "the only check on the system is corporate chutzpah"
8:50pm October 21st 2013 via Hootsuite in reply to
The shutdown and the DC stinkbug. WSJ http://t.co/gzKLkubrxh
8:50pm October 21st 2013 via Hootsuite
RT @drbachinsky: How much of your genome do you inherit from a particular grandparent? http://t.co/w6blL5v7Mq via gcbias blog
7:50pm October 21st 2013 via Hootsuite in reply to
RT @HistoryInPics: A fearless worker on the unfinished Golden Gate Bridge, 1935 http://t.co/tSkFP7GjBa
6:50pm October 21st 2013 via Hootsuite in reply to
RT @problogger: blogging is.... http://t.co/CfWDX0BWhG
5:50pm October 21st 2013 via Hootsuite in reply to
Wang: Because of the faster speed, lower cost, flexibility in throughput, Proton is well-suited for NIPT #ionworld2013
5:44pm October 21st 2013 via Hootsuite
Wang: For NIPT both platforms 100% consistent. 1500 samples tested (with unk. copy numbers) #ionworld2013
5:43pm October 21st 2013 via Hootsuite
Wang: Sensitivity / Specificity - a balance between the two. Comparing platforms HiSeq to Proton, similar numbers. #ionworld2013
Wang: 2K maternal plasma samples tested since June; pooling 10 samples / run (5M reads/sample). 99% fetal gender. #ionworld2013
5:42pm October 21st 2013 via Hootsuite
Wang: NIPT in China - 20M newborns/year there. T21, T13, T18 screens on Proton. #ionworld2013
5:41pm October 21st 2013 via Hootsuite
Wang: Has to do billions of long reads, nice to see preliminary Avalanche data (isothermal template prep in development) #ionworld2013
5:40pm October 21st 2013 via Hootsuite
