Mason: 'Whoever guesses the number of slides wins a patch'. (Worthy effort, the patch is really cool looking.) #AGBT17
2:35pm February 16th 2017 via Hootsuite
Mason: #AGBT17 shows chart from '13 ref https://t.co/56f6LMwKSM about the 'inevitable march' of epigenetic change and eventual death
2:34pm February 16th 2017 via Hootsuite
Mason: Epigenetic evolution - EPM '16 ref https://t.co/1BwtwuPQJn Shows figure from '05 of epigenetic marks shift in twins #AGBT17
2:33pm February 16th 2017 via Hootsuite
Chris Mason (Weill Cornell NY): DNA sequencing, genome assembly, and epigenetics on the international space station #AGBT17
2:31pm February 16th 2017 via Hootsuite
Q: Repetitive sequence improvements? Simpson: Trinucleotides easier than homopolymers, but it is HPs where the work is. #AGBT17
2:29pm February 16th 2017 via Hootsuite
Simpson: A2: there will be more work coming out with new motifs for modified bases soon. #AGBT17
2:28pm February 16th 2017 via Hootsuite
Q: Ability for novel bases? Simpson: Paper on Bioarxiv can pull out motif (via t-test) looking at raw data. They work model-based #AGBT17
2:27pm February 16th 2017 via Hootsuite
Q: For 1D^2 how often it gets to other side? Simpson: Has only seen data he's been given. #AGBT17
2:26pm February 16th 2017 via Hootsuite
Simpson: Concludes that scrappie basecaller is promising; as well as looking at long-range methylation patterns #AGBT17
2:25pm February 16th 2017 via Hootsuite
Simpson: Working on single-read, single-base methylation calling (haplotype-phased) with nanopolish tool #AGBT17
2:23pm February 16th 2017 via Hootsuite
Simpson: 5meC change signal in a detectable way. Did an experiment modeling C vs 5meC via Gaussian distribution #AGBT17
2:22pm February 16th 2017 via Hootsuite
1D^2 illustrated (too hard to capture!) RT @coregenomics: #agbt17 https://t.co/icAjHF79sr
2:20pm February 16th 2017 via Hootsuite
Simpson: New method of generating 2D reads, 1D^2: being able to pull in opposite strand without hairpin. Shows accuracy comparison #AGBT17
2:19pm February 16th 2017 via Hootsuite
RT @coregenomics: Variability in @nanopore yield by JS at #agbt17 https://t.co/dPZ11nicaN
2:17pm February 16th 2017 via Hootsuite
Simpson: Latest basecaller ("scrappie") estimates length based upon time measurement takes for sequence #AGBT17
Simpson: Homopolymers (HP) 'main source of residual errors in consensus sequences'. Early basecallers collapse HP's to 6-mers #AGBT17
2:16pm February 16th 2017 via Hootsuite
Simpson: Coming up with better human assembly, 'very much a work in progress' with a fraction of the data. https://t.co/CXSIOOlYAq #AGBT17
2:14pm February 16th 2017 via Hootsuite
Simpson: Accuracy of 1D reads was 82-83%. Canu assembly, NG50 3.0MB. 50-10 contigs across each chromosome #AGBT17
2:13pm February 16th 2017 via Hootsuite
Simpson: Ave readlength across sites, from 6.6kb, rapid library now have longer lengths #AGBT17
2:12pm February 16th 2017 via Hootsuite
Simpson: Optimize draft genome seq with HMM constraints; used for mobile Ebola using toookit called Nanopolish #AGBT17
2:11pm February 16th 2017 via Hootsuite
Simpson: RNN=recurrent neural network, shown improvement. They are interested in groups of nanopore reads, generating consensus #AGBT17
2:09pm February 16th 2017 via Hootsuite
Simpson: Uses speech recognition algorithms following similar path, from HMM to neural networks. R9 metrichor, nanonet, deepnano #AGBT17
2:08pm February 16th 2017 via Hootsuite
Simpson:Hidden Markov models were first-gen basecallers. Latest ones: moved to recurrent neural networks. #AGBT17
2:07pm February 16th 2017 via Hootsuite
Simpson: Signal-level analysis: what bases are registered by what current events? Existing tries to use 5- or 6-base calls #AGBT17
2:06pm February 16th 2017 via Hootsuite
Simpson: 0.5s, 60 picoA of current. Sees the next base, signal drops to 40 pA, then up to 50 etc. Stepwise changes #AGBT17
2:05pm February 16th 2017 via Hootsuite
Simpson: 4k samplings per second, as current is disrupted there are characteristics of the DNA as it passes thru #AGBT17
2:04pm February 16th 2017 via Hootsuite
Simpson: Shows a drawing of the protein pore, the charge differential across the membrane, and ssDNA able to pass through the pore #AGBT17
Simpson: Will intro nanopore, signal-level analysis, initial results, and direct detection of 5-meC #AGBT17
2:03pm February 16th 2017 via Hootsuite
Jared Simpson (OICR Canada) Nanopore sequencing of a human genome #AGBT17
2:02pm February 16th 2017 via Hootsuite
.@TorontoGenomics As time goes on, there will be fewer and fewer who will understand what this X-ray film business is all about!
2:00pm February 16th 2017 via Hootsuite in reply to TorontoGenomics
I wish I had photos of pets to share with #AGBT17. Here's a photo from circa 1993 - enjoy. https://t.co/TlQC6JrpRr
1:56pm February 16th 2017 via Hootsuite
Have another free-of-charge #AGBT17 single-cell RNA-seq reference https://t.co/MuYgtKA3S8
1:43pm February 16th 2017 via Hootsuite
Here's an interesting single-cell spatial gene expression paper that has nothing at all to do with #AGBT17. https://t.co/AgeKmXi1JX
1:40pm February 16th 2017 via Hootsuite
RT @RobAboukhalil: Great talk by Ido Amit on █████ using single-cell sequencing to explore the underlying biology of █████ █ ███ █ #agbt17
1:38pm February 16th 2017 via Hootsuite
Ido Amit (Weizmann Inst Israel) has asked for me not to tweet - so his loss (and ours) #AGBT17
1:33pm February 16th 2017 via Hootsuite
Livelli: Combined w/CRISPR for endogenous biology. Super-resolution microscopy uses HaloTags, single-molecule, single-cell imaging #AGBT17
11:48am February 16th 2017 via Hootsuite
Livelli: Can attenuate affinity between components of donor-acceptor molecules. Thus 2nd prot-prot interaction NanoBIT method #AGBT17
11:47am February 16th 2017 via Hootsuite
Livelli: Able to do kinome-wide profiling w/ a library of NanoLuc-kinase fusions. Suggests it could be used for pers med #AGBT17
11:46am February 16th 2017 via Hootsuite
Livelli: Developed an in-cell target / ligand system called NanoBRET Info here: https://t.co/CyzLQxBfdP #AGBT17
11:44am February 16th 2017 via Hootsuite
Livelli: Able to do directed evolution to increase light output greatly, Donor protein; getting a HaloTag fluorescent acceptor #AGBT17
11:43am February 16th 2017 via Hootsuite
Livelli: Protein degradation, abundance, localization, reporting techology. Oplophorus lucferase - https://t.co/IYu6Qnu6CJ #AGBT17
11:42am February 16th 2017 via Hootsuite
Livelli: Emphasizing the functionalization of genomic info; interrogation of drug/target; molecular sensors; protein interaction #AGBT17
11:40am February 16th 2017 via Hootsuite
RT @Supergecko: @iontorrent @thermofisher presenting S5 kits: 600bp seq kit available, 800bp seq kit in development #agbt17
11:39am February 16th 2017 via Hootsuite
Thomas Livelli (Promega WI) Next generation integrative life science research #AGBT17
Ellis: Shows data from PCR-free from 250ng, feels they can go down as little as 50ng or 25ng at scale 'for the greedy X10's' #AGBT17
9:53am February 16th 2017 via Hootsuite
Ellis: Shows GC curves, answering one question, and automation data for 96-well. Next step is to 384-well. PCR-free has 1ug input #AGBT17
9:52am February 16th 2017 via Hootsuite
Ellis: Next challenge is scaling, unsure the effect of different contaminants on a new fragmentation method, effect on GC coverage #AGBT17
9:50am February 16th 2017 via Hootsuite
Ellis: Have settled on a method, cancer group has indicated quality of data is excellent, similar to bulk DNA cp to LCM #AGBT17
9:47am February 16th 2017 via Hootsuite
Ellis: Duplicate rate held way down - only 30% below 1ng input. LCM to purification still a problem. Staining, buffers, proteases #AGBT17
9:46am February 16th 2017 via Hootsuite
Ellis: Found 10x-15x more library yield compared to other methods; a proxy for complexity. Same complexity graph - down to below 1ng #AGBT17
9:45am February 16th 2017 via Hootsuite
