dib-lab/2015-09-10-scihack

Inaugural Group Science Coding Session

2015-09-10-scihack's Issues

use case example

1. Run build-sbt.py with parameters suggested in #8

• takes list of samples, raw reads
• each leaf in tree are kmers in whole sample
• output to disk as .json tree structure, false positive rates (hidden folder don't need to look in hash tables - will need to delete if run a bunch of times...)

if 0.5 false positive, need to make bigger

--print-tree

(see build-sbt.py --help)

1. find-false-isoforms.py

• run with output from 1. and assembly
• outputs .good - those that meet threshold in at least one leaf - and .bad - false isoforms

90% of kmers in the transcript need to show up in one of the samples
if not, then suggests that false isoforms

need way to validate results, take mm9 RNAs

Mus musculus reference cDNAs

ftp://ftp.ensembl.org/pub/release-75/fasta/mus_musculus/cdna/