[dib-training] [saviran at ucdavis.edu: [gc-structure-function] Structural and functional genomics focus group meeting this Friday at 2pm]

[C. Titus Brown]

This might be of interest more generally, for those who are doing RNAseq.

best,
--titus

----- Forwarded message from Sharon Aviran <saviran at ucdavis.edu> -----

From: Sharon Aviran <saviran at ucdavis.edu>
Date: Mon, 8 Feb 2016 22:33:37 -0800
Cc: Harold Pimentel <haroldpimentel at gmail.com>,
	Nicolas Bray <nicolas.bray at gmail.com>
To: gc-structure-function at ucdavis.edu
X-Mailer: Apple Mail (2.2098)
Subject: [gc-structure-function] Structural and functional genomics focus
	group meeting this Friday at 2pm

Hi all,

The structural and functional genomics focus group will be meeting this Friday at 2pm. This time, we will host 2 guest speakers, Dr. Nick Bray and Harold Pimentel, who will cover two related projects on rapid gene expression quantification and differential expression analysis from RNA-Seq data. 

Note that the location is room 2202 in GBSF. 
 
Title: Ultrafast RNA-seq Analysis with Kallisto and Sleuth???
Speakers: Nicolas Bray and Harold Pimentel, Innovative Genomics Initiative (IGI) and the Computer Science Department, UC Berkeley (respectively)
Where: GBSF 2202
When: Friday, Feb. 12, at 2pm
Coffee and pastries will be provided. 

We hope to see you there. 
Sharon


Abstract:

While the continued expansion in humanity's sequencing capability is typically measured in terms of genomes sequenced, we have also entered an era in which RNA-seq experiments involving hundreds of samples are becoming routine. Traditional statistical analysis of such data involves hours of computation time first aligning reads and then analyzing the resulting alignments in order to estimate the abundances of the various transcripts present.

Nicolas Bray (kallisto)
Here we will present a method, implemented in the program kallisto, which is capable of progressing from raw sequencing reads to transcript abundances for an average RNA-seq experiment in under five minutes on a standard laptop while preserving the accuracy of traditional analyses. This is accomplished by approximately computing what we call "pseudoalignments" of the reads, giving sufficient information for a minimal model of RNA-seq.

Harold Pimentel (sleuth)
The use of this minimal model also allows for rapid bootstrap analysis to estimate the uncertainty in the resulting abundances. These uncertainty estimates enable new differential expression modeling approaches where the variance can be decomposed into abundance estimation variability and biological variability. This decomposition results in fewer false positives in differential expression analysis when abundance estimation variability is high, while maintaining similar accuracy when abundance estimation variability is low.

This is joint work with Pall Melsted and Lior Pachter.


Sharon Aviran, Ph.D.
Assistant Professor
Biomedical Engineering Department
and Genome Center, UC Davis
451 Health Sciences Dr.
Genome and Biomedical Sciences Building (GBSF) 2319
Davis, CA 95616


----- End forwarded message -----

-- 
C. Titus Brown, ctbrown at ucdavis.edu