Trinity/Rsem/Edger Pipeline...Now What?

Question

[Biogeek]

HI people,

I am a bit new to all of the Bioinformatics but so far I have produced data from the mentioned Trinity/RSEM/edgeR pipeline...

I basically have 3 sample condtions -with 3 replicates each.

Control,Low and High conditions.

I have my fpkm file with expression values and transcript id (e.g Comp1234_seq1)

Now I have the Trinity. fasta file and I'm wanting to know how I can BLAST analysis multiple sequences at a time(talking a few hundred).

I have compared:

Genes and Isoforms present in Control only Genes and Isoforms present in Low only Genes and Isoforms present in High only

I now have excel spreadsheets with only the genes/Isoforms present in these conditions with transcript id and expression value... I want to BLAST these against RNA-seq databases. Is there a program which lets me merge the filtered sequences with my Trinity.fasta file or am I thinking off track and is there another way to simply do this?

Also..

How can I also get the top 1000 increased transcripts between samples? How can I get top1000 decreased transcripts between samples? The least changed transcript expression over all samples?

Any help is greatly appreciated. Many thanks :)

[Floris Brenk]

Did you checked out the downstreams analysis options of Trinity? http://trinityrnaseq.sourceforge.net/#Downstream_analyses and especially http://trinotate.sourceforge.net/?

Answer 1

[Charles Warden]

A couple suggestions:

1) You can see this discussion on the differential expression analysis. I think it is actually pretty tricky because of difficulties in clearly mapping transcripts between samples:

A: Differential expression analysis using two different de novo assembled transcriptomes

2) I've actually found that normal de novo assembly seemed to produce better results than the programs that were supposed to be specific for RNA-Seq (namely, CLC Bio de novo seemed to work best with the RNA-Seq de novo assembly). I also thought that Oases/Velvet worked better than Trinity.