Protocols (workflow) for differential gene expression analysis of de novo assembled transcriptomes

Question

[lidanqing1990]

Dear all,

I am new for bioinformatics and feel very confusing about differential gene expression analysis for de novo assembled transcriptomes. I knew Tophat +Cufflinks or HISAT+StringTie+Ballgown could be used for differential gene expression analysis for transcriptomes with reference genome. So how can I do differential gene expression analysis for de novo assembled transcriptomes? May I use the same workflow with transcriptomes with reference genome? A friend told me that I could use Bowtie + DESeq to do that, is there any other choice?

Can anyone give me suggestions? Thank you so much.

Answer 1

[predeus]

Most of the mentioned software tools are dramatically outdated -

Please stop using Tophat https://t.co/Es4ohxOEyx Cole and I developed the method in *2008*. It was greatly improved in TopHat2 then HISAT & HISAT2. There is no reason to use it anymore. I have been saying this for years yet it has more citations this year than last #methodsmatter

— Lior Pachter (@lpachter) December 2, 2017

You have to solve one more problem than with normal differential expression analysis

Normally, you want to 1) quantify the transcripts/genes; 2) do differential expression analysis. A good quantifier is RSEM/kallisto/salmon. A good differential expression package is limma-voom/edgeR/DESeq2.

With de-novo transcripts, you have additional problem of falsely differentially expressed transcripts that actually belong to the same gene.

People suggest Corset and Lace to handle the latter problem.