Mao:Decided to not report incidental findings - parents are vulnerable with a sick child. #NGDx15
9:54am August 18th 2015 via Hootsuite
Mao: Selection of pts - 'high acuity', making a Dx time-critical, known genes that cause symptoms, and other testing $$ #NGDx15
9:50am August 18th 2015 via Hootsuite
Mao: NICU library prep 1d, cluster + seq 35h, data analysis 1-4d. >99% coverage at >15x #NGDx15
9:49am August 18th 2015 via Hootsuite
Mao: Less expensive than WES. Missing deep intronic regions, 5', 3' UTRs, repetitive and pseudogenes. Not including tri-nt repeats #NGDx15
9:48am August 18th 2015 via Hootsuite
Mao: 4,503 HGMD genes for inherited diseases; 98% overlaps OMIM 4K known disease-causing genes. 15Mb design. #NGDx15
9:47am August 18th 2015 via Hootsuite
Mao: NICU needs quick and precise Dx, as neonatal presentation is atypical. With precise Dx, many diseases are treatable #NGDx15
9:46am August 18th 2015 via Hootsuite
Mao: These numbers do not include expensive tests. #NGDx15
9:45am August 18th 2015 via Hootsuite
Mao: Understood to be a number of genetic and environmental factors. High mortality. Cost/day is $1,379, median stay 35d #NGDx15
9:44am August 18th 2015 via Hootsuite
Mao: Lists the most common disorders in NICU. 150K babies in the US born with birth defects. 3/100 have some major birth defect #NGDx15
9:43am August 18th 2015 via Hootsuite
Rong Mao, ARUP (Salt Lake City, UT) "Rapid turnaround inherited disease panel in the NICU" #NGDx15
Arnold:Q:Homopolymers? A:He's very familiar with the problem, and took that into consideration in the assay design #NGDx15
9:40am August 18th 2015 via Hootsuite
Arnold:Showed interesting depth vs. samples per chip plot; 14/chip hit a >200x-300x coverage. RNA-Seq - 35M reads, >90% align #NGDx15
9:39am August 18th 2015 via Hootsuite
Arnold: Shows Ion Proton summary - very high uniformity, very high coverage. Mean depth >150x. Mapped reads typically >2M/smpl #NGDx15
9:37am August 18th 2015 via Hootsuite
Arnold: Assessing workflow and performance: Used >15y old samples, Ashkenazi BRCA1/2 variants, on ASO dot blot #NGDx15
9:36am August 18th 2015 via Hootsuite
Arnold: @IcahnInstitute rank among the largest computational systems in academic medicine #NGDx15
9:35am August 18th 2015 via Hootsuite
Arnold: Data: 160-224Gb total filtered bases per day. Use Ion Reporter for var calling, annotation. Sequencing at 50% capacity #NGDx15
9:34am August 18th 2015 via Hootsuite
Arnold: 8 Chefs support 8 Protons; a Chef takes 14h for two Ion PI chips. 14 barcoded samples / chip. Plan to get add'l systems #NGDx15
9:32am August 18th 2015 via Hootsuite
Arnold: Use Equalizer (from @servingscience ) for normalizing after library prep. The site is a combination of research and clinical #NGDx15
9:31am August 18th 2015 via Hootsuite
Arnold: Functional QC via Fluidigm SNP type fingerprint assay, for AIM (ancestry informative markers) and gender SNPs #NGDx15
9:30am August 18th 2015 via Hootsuite
Arnold: BioMcroLab VolumeCheck (x2) for elution volumes; Advanced Analytical CE fragment analyzer for library sizing. #NGDx15
9:29am August 18th 2015 via Hootsuite
Arnold: Analytical QC - use Trinean DropSense 96 (x2) with cDrop for NA quant, purity. Looks for specific contaminant carryover #NGDx15
9:28am August 18th 2015 via Hootsuite
Arnold: Their facility can do 750 samples/day, Hamilton/Chemagic Star, 3x QIAsymphony, lots of automation #NGDx15
9:27am August 18th 2015 via Hootsuite
Arnold: (Dudley YouTube video on Resilience here: http://t.co/h9dge1xCuq ) #NGDx15
Arnold: More about @icahninstitute Resilience Project here: http://t.co/la5BCCcGVC An opportunity for people to help. #NGDx15
9:24am August 18th 2015 via Hootsuite
Arnold: Covers the Resilience Project - the retrospective study with 600K individuals. ID'd 13 potential resilient people #NGDx15
9:23am August 18th 2015 via Hootsuite
Arnold: Reviews AmpliSeq tech, their SuperPanel uses 12.5K primer pairs/pool and 2 pools. Uses Ion Chef automation. #NGDx15
9:21am August 18th 2015 via Hootsuite
Arnold: ~25K targets, optimized for spec and sens. Points out how easy AmpliSeq Designer is for modifying panels to meet needs #NGDx15
9:20am August 18th 2015 via Hootsuite
Arnold: Everything now is RUO. Using @iontorrent AmpliSeq and RNA-Seq, can use blood, saliva, tissue. SuperPanel: >700genes #NGDx15
Arnold: Submitting CLIA/CAP later this mo. 50K samples/y. 31K samples in-house. Support with a lot of data with a res 'super-panel' #NGDx15
9:19am August 18th 2015 via Hootsuite
Arnold: Why CT: economical, space to expand, highly trained workforce. (Not mentioned - building where GS20 was invented/developed) #NGDx15
9:17am August 18th 2015 via Hootsuite
Arnold: Mt. Sinai in Branford CT - will talk about goals, projects, and workflow #NGDx15
Todd Arnold (Icahn School Med, Mt. Sinai) "Considerations for optimizing results with @iontorrent Proton Sequencing" #NGDx15
9:16am August 18th 2015 via Hootsuite
Lin:Q:How will methylation be used in future? A:Key term - actionability. Now only preparing for the future where evidence is clear #NGDx15
9:13am August 18th 2015 via Hootsuite
Lin: Q:RNA-seq from FFPE? A:They look for fusions, mainly.Would love to do RNA-Seq from fresh tissue, but PDX is the priority #NGDx15
9:12am August 18th 2015 via Hootsuite
Lin: Germline reporting 'will become more and more important' as more cancer exomes are sequenced. #NGDx15
9:10am August 18th 2015 via Hootsuite
Lin: Germline - works with Les Biesecker (NHGRI), for incidental finding. Req 'at least 3 papers' for discussion before sign-out #NGDx15
9:09am August 18th 2015 via Hootsuite
Lin: Germline 'we stick to ACMG genes', conservative on what is reported out. #NGDx15
9:08am August 18th 2015 via Hootsuite
Lin: A constant clinical validation. Build internal db to compare back to 'is essential' (@seracare Seraseq NGS reference material) #NGDx15
Lin: Clinical / regulatory: 'higher than CLIA, little lower than NY'. Choose validation samples, QC, QA, proficiency, training #NGDx15
9:06am August 18th 2015 via Hootsuite
Lin: Very conservative regarding annotation and implications for clinical care, access to many resources. #NGDx15
9:05am August 18th 2015 via Hootsuite
Lin: Somatic SNV via BWA-mem, GATC, Mutect. Somatic index via Strelka. Medium/large indels are still hard for high sens. #NGDx15
9:04am August 18th 2015 via Hootsuite
Lin: Somatic clalers: Strelka, Mutect, Varsan, Lofreq - here is where the most development and ongoing work is being done #NGDx15
9:02am August 18th 2015 via Hootsuite
Lin: Bioinformatics: aligners: BWA, BWA-mem, Novoalign. Var callers: Unified genotyper, haplotype caller, Platypus, BAM2MPG #NGDx15
Lin: IT system - highly backed-up. FASTQ on Biowulf. BAM is intermediate; also VCFs. Looking into more efficient compression #NGDx15
8:59am August 18th 2015 via Hootsuite
Lin: WES Specificity: 99.99% across ('a funny metric to calculate') but sens is 95.6% - 98.5%; PPV 89% - 98% #NGDx15
8:58am August 18th 2015 via Hootsuite
Lin: For WES, goal is 90x, median 124x coverage. Compared to MyCancerGenome 191 hotspots, ave coverage 459x, 11 expts #NGDx15
8:57am August 18th 2015 via Hootsuite
Lin: Targeted tumor coverage median coverage 346x (goal 300x); germline goal is 150x #NGDx15
8:56am August 18th 2015 via Hootsuite
Lin: Whole-exon is 564 genes/9442 exons; whole-genes (6); PGx (1906 markers) #NGDx15
8:55am August 18th 2015 via Hootsuite
Lin: Two panels - one for mutations, another for CNV and fusions. Design - listed a large number of existing pane. NCI, @iontorrent #NGDx15
8:54am August 18th 2015 via Hootsuite
Lin: Wet lab: did WES comparison between SureSelect, SeqCap EZ - fold-coverage by % of exome covered. Comparable but SureSelect won #NGDx15
8:53am August 18th 2015 via Hootsuite
