January 2017

Robert Green (Harvard Med, also director of https://t.co/6iSHdGRyFW) Evidence of clin util, cost-benefit and prof rec's and reimb #PMWC17

12:14pm January 25th 2017 via Twitter Web Client

Scott: Genome Medical https://t.co/zY6HzARm9R staffed with doctors, to give guidance to consumers, consulting to physicians too. #PMWC17

12:12pm January 25th 2017 via Twitter Web Client

Scott: Medical dx grade to rule in or out for high-risk, or health screen? "There's an enormous amount of naivete" #PMWC17

12:10pm January 25th 2017 via Twitter Web Client

Scott: Cost - from $1K to $25K? Whole-genome or Hole-genome? Gaps at 30x coverage; or targeted? #PMWC17

12:10pm January 25th 2017 via Twitter Web Client

Scott: Sequencing your genome isn't an event: it is a lifelong journey, but where do you start? Purpose: medical? Consumer? #PMWC17

12:09pm January 25th 2017 via Twitter Web Client

Scott: Shows figure how rare is actually relatively common - eg PGx traits. We have so little knowledge of genetic predisposition. #PMWC17

12:07pm January 25th 2017 via Twitter Web Client

Scott: Overlays annual cellphone subscribers and total number of genomes sequenced. 'We're in very early days' #PMWC17

12:06pm January 25th 2017 via Twitter Web Client

Randy Scott (Invitae) Genetic Testing in Healthy Populations #PMWC17 Exciting but not in a short timeframe: a 'last-mile problem'

12:03pm January 25th 2017 via Twitter Web Client

Iafrate: CNV analysis from cfDNA - a ratio of 1.15x to normal. 1.8% TP53; 15x of HER2, in 0.018 of cells = 0.982/2 (2copies) = 1.15 #PMWC17

11:55am January 25th 2017 via Twitter Web Client

Iafrate: Onto gene fusions, looking at cfRNA in plasma. qPCR for EGFR at 52bp amplicon length, can't pull it out as of yet. #PMWC17

11:53am January 25th 2017 via Twitter Web Client

Iafrate: Resistance - T790M, C797S showed figure from Janne NEJM '15 https://t.co/kS7rSD8Zdd #PMWC17

11:51am January 25th 2017 via Twitter Web Client

Iafrate: Going to TCGA, 77% lung adenocarcinoma would have one mut in their current panel #PMWC17

11:49am January 25th 2017 via Twitter Web Client

Iafrate: No concordance between cfDNA mutation allele freq ng/mL varied 10x to 180x No clear understanding, may also be pre-analytic #PMWC17

11:47am January 25th 2017 via Twitter Web Client

Iafrate: 10 pts w/muts had active clinical progression at blood draw. 3/27 mut-neg were in progression, had muts not on panel #PMWC17

11:46am January 25th 2017 via Twitter Web Client

Iafrate: 200 GSP sets, 28 genes, solid tumors. EGFR, KRAS, PIK3CA etc. Missed 0.1% picked up 0.13%, 50ng input, expected 0.3% sens #PMWC17

11:44am January 25th 2017 via Twitter Web Client

Iafrate: Urine: Genitourinary tract tumor residual disease/predictive. Early det: very hard, TP53 #PMWC17

11:42am January 25th 2017 via Twitter Web Client

Iafrate: CSF avoid brain biopsy (TERT, TP53, IDH1); cyst fluids/bile duct/pleural fluid #PMWC17

11:41am January 25th 2017 via Twitter Web Client

Iafrate: Breast initial Dx, resist (PIK3CA, ESR1); melanoma inital DX, resis (BRAF, NRAS, HLAA) #PMWC17

11:41am January 25th 2017 via Twitter Web Client

Iafrate: Common clin uses: likely multiple tissue-spec panels due to depth needed. Lung initial Dx, resist (EGFR, KRAS, ALK, ROS1) #PMWC17

11:40am January 25th 2017 via Twitter Web Client

Iafrate: Appplications: MRD, Resistance, initial genotyping when no tissue avail, early det/screening, research on heterogeneity #PMWC17

11:39am January 25th 2017 via Twitter Web Client

Iafrate: Also sarcoma fusion (ArcherDx), hematologic tumor (ILMN), heme tumor fusion (ArcherDx) #PMWC17

11:38am January 25th 2017 via Twitter Web Client

Iafrate: The idea was to democratize ctDNA sequencing, to smaller clinical dx labs. At MGH, use AMP for solid tumor, fusion RNA... #PMWC17

11:37am January 25th 2017 via Twitter Web Client

John Iafrate (Mass Gen Hosp) Cell-free DNA sequencing: clinical utility #PMWC17

11:36am January 25th 2017 via Twitter Web Client

Le: Showed true variant - how to achieve that Q60 - shows true positives across 7/8 control samples. (Unclear what AF, likely 0.1%) #PMWC17

11:35am January 25th 2017 via Twitter Web Client

Le: Conclude need 40x coverage; showed 95% confidence sensitivity, background noise, measured AF #PMWC17

11:33am January 25th 2017 via Twitter Web Client

Le: One trick: modeling a binomial distribution for 'alternate observations' (AO) with plot of probability where AO's is noise #PMWC17

11:31am January 25th 2017 via Twitter Web Client

Le: MBC (molecular bar code) error correction diagram of noise elimination shown graphically. #PMWC17

11:30am January 25th 2017 via Twitter Web Client

Le: Credits Vogelstein with the first barcoding method Safe-Seq, to eliminate FP's with error correction. #PMWC17

11:29am January 25th 2017 via Twitter Web Client

Le: Q30 regular metric, many false positives exist. Error correction needed. Systematic sample storage/fixation, sample prep, PCR #PMWC17

11:28am January 25th 2017 via Twitter Web Client

Le: Higher depth, higher noise. Q4 means 1 true var, 10 reads w/error. PPV of 11%. Need Q60 to get 100% PPV. #PMWC17

11:26am January 25th 2017 via Twitter Web Client

Le: 1000 molecules in, 1000 molecules seq? Q30 provenance, Phred w/HGP. 1000 reads at Q30, 100bp: 1 read in 100 will be a true var #PMWC17

11:25am January 25th 2017 via Twitter Web Client

Le: Avoid GIGO (garbage in...) with adequate sample handling, avoid loss of library diversity, adequate sequencing sampling (depth) #PMWC17

11:23am January 25th 2017 via Twitter Web Client

Le: Need adequate sampling of ctDNA molecules; and preservation of library diversity. A debate - 200cc's of blood instead of 20cc? #PMWC17

11:21am January 25th 2017 via Twitter Web Client

Le: Early det is of high interest, also for resistance monitoring, both need very high sensitivity for very low level detection #PMWC17

11:20am January 25th 2017 via Twitter Web Client

Long Le (Mass Gen Hosp) The dist of individ. molecules of ctDNA from Bettegowda et al doesn't represent the numbers of normal cfDNA #PMWC17

11:18am January 25th 2017 via Twitter Web Client

Myers: Trying to detect var's at or near the limit of sequencing error rate. #PMWC17

11:16am January 25th 2017 via Twitter Web Client

Myers: Opposing primers may only capture 11% of unique molecules (due to redundant clones) AMP is spec for ctDNA #PMWC17

11:14am January 25th 2017 via Twitter Web Client

Myers: Spent 18 mo on ctDNA: improved capture efficiency by looking at independent observation (i.e. unique start sites) #PMWC17

11:12am January 25th 2017 via Twitter Web Client

Myers: AMP is a platform - measuring fusions, splicing, CNVs, expression, SNVs, indels. Chemistry, workflow, analysis software #PMWC17

11:11am January 25th 2017 via Twitter Web Client

Myers: Adapters are half-functional after ligation; two rounds of PCR (universal once side, gene-specific nested on other) #PMWC17

11:10am January 25th 2017 via Twitter Web Client

Myers: Points to '14 Nat Med ref https://t.co/ONPcxJo93P describing Anchored Multiplex PCR (AMP) #PMWC17

11:09am January 25th 2017 via Twitter Web Client

Myers: Need scale for simultaneous measurement of 1000's of markers; need digital, wide dynamic range: both provided by NGS #PMWC17

11:08am January 25th 2017 via Twitter Web Client

Myers: Many tumor types shed ctDNA; both healthy cfDNA and ctDNA exist in plasma #PMWC17

11:07am January 25th 2017 via Twitter Web Client

Myers: Monitoring treatment via ctDNA eliminates problem of going in for a 2nd biopsy; fig from https://t.co/RuQoIHUZ48 #PMWC17

11:06am January 25th 2017 via Twitter Web Client

Myers: Resistance mutations, secondary drivers: after initial Dx, first-line targeted; then monitor, give a second-line Rx #PMWC17

11:04am January 25th 2017 via Twitter Web Client

Myers: Using lung cancer pre-1970 it was viewed as a single disease; now molecularly characterized. #PMWC17

11:03am January 25th 2017 via Twitter Web Client

Jason Myers (ArcherDx) Next-generation Liquid Biopsy Testing - Promise and Limitations #PMWC17

11:01am January 25th 2017 via Twitter Web Client