Starcevic: Sensitivity calculation is from 47 known positive controls; FDR uses http://t.co/wAytUNhA2y consortium regions #gensoltour
1:55pm September 30th 2014 via Twitter Web Client
Starcevic: 0.4% calls of a novel alt allele w/known dbSNP - clearly an error in dbSNP #gensoltour
Starcevic: 47 positive controls had difficult indels; 97.6% calls concordant; 2.0% were novel alleles (nothing in dbSNP) #gensoltour
1:54pm September 30th 2014 via Twitter Web Client
Starcevic: Resilience panel - >4K disease-ass'd regions, 4.4MB targeted in 26K amplicons. Pilot: 47 positive ctls, 199 neg #gensoltour
Starcevic: Resilience project (reviewed on Behind the Bench http://t.co/0FUI0qlBAH ) pilot >700 disease ass'd genes #gensoltour
Starcevic: Points out the need for HTP NGS, and are about to open up a CT laboratory. From sample collection thru analysis #gensoltour
Starcevic: Also v. large bioinformatics infrastructure. Many research programs, heavy focus on cancer #gensoltour
Starcevic: Involves EHR, population genetics, multi-platform (both NGS and microarray-based) liquid handling automation #gensoltour
Starcevic: The Icahn Inst - ~300 research scientists, work on translation; ultimate goal is personalized medicine #gensoltour
1:53pm September 30th 2014 via Twitter Web Client
Next: Daniela Starcevic, Icahn Sch Medicine Mt. Sinai “Developing targeted gene panels for clinical research applications” #gensoltour
Broeckel:Q:Ligation of miRNA and bias in lib. prep? A:Pointed to PLoS One manuscript, used size selection, use of IVT ctrl #gensoltour
Broeckel:Q: serum miRNAs in exosomes? A:Yes they have to be isolated first #gensoltour
Broeckel: "Unprecedented platform for functional genome analysis"; multiple RNA analysis platforms tailored to spec needs #gensoltour
Broeckel: Log2-fold change distribution also very similar results between RNA-Seq and AmpliSeq Whole Transcriptome #gensoltour
Broeckel: Showed distribution log10 plots of RPKM between RNA-Seq and AmpliSeq Whole Trancriptome - very close #gensoltour
1:52pm September 30th 2014 via Twitter Web Client
Broeckel: That is, intra-individual related from data clusters together. Difference between individual lines greater #gensoltour
Broeckel:Pointed out non-correlation between individuals (reassuring); clustered across samples shows stim/unstim similarity #gensoltour
Broeckel: For their iPSC's, >56% target detected at >10x. Highlights ERCC controls, nice linearity, R^2 >0.95 #gensoltour
Broeckel: >93% on-target, >65% target detected; still digging through analysis; showed UHR and HBR reference RNA data; #gensoltour
Broeckel: "V. straightforward and worked in our hands right from the beginning" regarding AmpliSeq Whole Transcriptome #gensoltour
1:51pm September 30th 2014 via Twitter Web Client
Broeckel: Pros: faster, pool multiple samples, reduced library prep/analysis time, cost. Cons: no diff. isoforms, no RNA-editing #gensoltour
Broeckel: 20.8K amplicons, high concordance. (For more details: Behind the Bench blog post http://t.co/5iNp2FSPXu ) #gensoltour
Broeckel: Moving to expression analysis after knockdown; now highlights work with AmpliSeq RNA whole-transcriptome #gensoltour
Broeckel: Knock-down exp.'s clear reduction in disease phenotype; consistent with fn; reduced BNP exp when ET-1 stim. #gensoltour
Broeckel: Known 'hypertrophy' miRNAs, qPCR confirmed; developed their own set of synthetic miRNA controls #gensoltour
1:50pm September 30th 2014 via Twitter Web Client
Broeckel: Their work recently published for details: PLoS One Sept 2014 http://t.co/4uLs93m1MW #gensoltour
Live shot from the #gensoltour today. http://t.co/gk2i44jAzQ
Broeckel: Many of their discovered miRNAs were not characterized in mIRbase; pipeline used SHRiMP2, miRDeep2, HTSeq, DESeq #gensoltour
1:49pm September 30th 2014 via Twitter Web Client
Broeckel: In the beginning (on Ion PGM & 318) looked at miRNA and control of expression; Refers to http://t.co/TJw8ZxP0Sy #gensoltou
Broeckel: Dose-response, chronic stimulation, shows phenotypic variation in fold-changes of specific genes #gensoltour
1:48pm September 30th 2014 via Hootsuite
Broeckel: Developed a production phenotyping workflow: each (250 lines) stimulated, RNA and protein harvested, qPCR screen #gensoltour
Broeckel: Comparing iPSC-derived cardiomyocytes with biopsy samples, comfortable with the cell model #gensoltour
Broeckel: Stimulated with inducers of hypertrophy, did gene exp. analysis and PCA; nice separation #gensoltour
1:47pm September 30th 2014 via Hootsuite
Broeckel: Need is for a better model. NHLBI's HyperGen; description here http://t.co/1G7o35d1C0 250 iPSC lines they produced #gensoltour
Broeckel: Left VH: indep CV risk factor, inter-individ. var., "hypertensive & diabetic heart disease" #gensoltour
Broeckel: ...to functional 'what the genome is doing'. Left ventricular hypertrophy, increase in mass of left side of heart #gensoltour
Broeckel: Begins with a 'Congratulations!' slide - advances in genetics / genomics. It's time to move from statistical ass'n... #gensoltour
1:46pm September 30th 2014 via Hootsuite
Next: Ulrich Boeckel, Medical Coll WI "RNA Sequencing research applications in human induced pluripotent stem cell tissues" #gensoltour
O’Connor:Q:Other protocols for 10K cells? Comment? A:May not be realistic for their work on clinical research samples #gensoltour
O’Connor:A: There are some 50 epigenetic-based drugs in various phases of testing #gensoltour
O’Connor:Q:How far along for drug dev.? A:Emphasizes the epigenetic component for disease, and for populations #gensoltour
1:45pm September 30th 2014 via Hootsuite
O’Connor: Being able to look at env. effects on phenotype, then examine via ChIP. Found Th1 protein. Potential v. exciting #gensoltour
O’Connor: Looking at T-cell activity, looking at STAT3, 4, FOXP3 genes and H3K27me3 peaks (repressive mark) #gensoltour
O’Connor: Also looking at a lung disease Sarcoidosis BAL, look at epigenetics of CD4+ T-cells. BAL: lavage of bronchus #gensoltour
O’Connor: In future: look at known asthma pathways with epigenetic inhibitors, look at changes in phenotype #gensoltour
O’Connor: Called out IRF4 transcription factor consistency; listed top hits, referred to Lugt et al 2014 Nature Immunol #gensoltour
O’Connor: Complex Venn: vast majority common across 5M, 3M (>92%) 1M (80%), their Q>=20 #gensoltour
O’Connor: Conf. (log-scale) values dropped from 32 (initially) now to 14 at 3M, 5M cells; 1M now at 17 'x-value' #gensoltour
O’Connor: Looking at peak overlap and p-values (confidence) w/different samples: 1M may need slightly higher confidence #gensoltour
1:44pm September 30th 2014 via Hootsuite
O’Connor: Will be helping @iontorrent produce a ChIP-Seq application note soon. Run report down to 1M cells shown - reliable #gensoltour
