cross species RNA-seq/orthologs

KateShepard · 04-15-2016, 07:37 AM

Dear Forum,

I am completely new to RNA-seq analysis so would be very appreciative of any advice regarding my query.

I have multiple RNA-seq samples from two closely related species, and would like to identify differentially expressed orthologous genes.

Is a correct approach to first identify orthologs between the two species using OrthoMCL, annotate those orthologs using (?), align all RNA-seq samples to the orthologs using bowtie2/tophat2, and assemble transcripts and identify differentially expressed transcripts using cufflinks and cuffdiff?

Thank you for your help and advice!
Kate

colindaven · 05-19-2016, 04:27 AM

Do you have a genome ? Where do you get your sequences from if not, transcript de novo assemblies ?

I would suggest proteinortho instead of orthomcl, as it is simpler in my experience.

I would use STAR instead of tophat2 if possible, featureCounts for read counting and edgeR/DEseq2 for DE analysis if it all works out.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
cross-species data - questions about normalization	trelek2	RNA Sequencing	1	05-23-2013 08:29 AM
cross species RNAseq	sidderb	Bioinformatics	0	04-11-2012 03:12 AM
Best Cross species aligner	Khanjan	Bioinformatics	0	02-07-2011 07:59 AM

04-15-2016, 07:37 AM	#1
KateShepard Junior Member Location: UK Join Date: Apr 2016 Posts: 5	cross species RNA-seq/orthologs Dear Forum, I am completely new to RNA-seq analysis so would be very appreciative of any advice regarding my query. I have multiple RNA-seq samples from two closely related species, and would like to identify differentially expressed orthologous genes. Is a correct approach to first identify orthologs between the two species using OrthoMCL, annotate those orthologs using (?), align all RNA-seq samples to the orthologs using bowtie2/tophat2, and assemble transcripts and identify differentially expressed transcripts using cufflinks and cuffdiff? Thank you for your help and advice! Kate

05-19-2016, 04:27 AM	#2
colindaven Senior Member Location: Germany Join Date: Oct 2008 Posts: 415	Do you have a genome ? Where do you get your sequences from if not, transcript de novo assemblies ? I would suggest proteinortho instead of orthomcl, as it is simpler in my experience. I would use STAR instead of tophat2 if possible, featureCounts for read counting and edgeR/DEseq2 for DE analysis if it all works out.