I am planing to find out differential gene expression between two samples (infected and non infected RNA-seq data of avian species from 454 GSFLX+ PLATFORM), but there is no reference genome available. So I am planning to perform a De No transcriptome assembly with Newbler. How can I determine the number reads assembled to particular gene (I have only contigs from the assembly) for statistical significance testing. Are any normalization procedures required? Please suggest some common software for this analysis.
Thanks
•
written
by
nayaganviji •
When you don't have a reference genome, makes sense to assembly your sequences to get the transcripts, one strategy I have seen before is to assemble all your libraries in one run to get the better representation of the transcriptome and then map back your reads per library to get the counts per condition. Then you can use R:Bioconductor edgeR or DESeq to deal with normalization and differentially expressed transcripts.
This is what Brian Haas recommends when he teaches. The Trinity folks put together some documentation on how to do it with Trinity here. You don't have to use Trinity and EdgeR but the logic is all laid out.
http://trinityrnaseq.sourceforge.net/analysis/diff_expression_analysis.html
Log in to answer this question.
Have you searched the forum? A similar question has been asked: http://www.biostars.org/post/show/17611/ Also, are you familiar with R and bioconductor? There are some powerful tools between those two that can answer the questions you are asking.