Using Trinity for RNA-Seq analysis in model organisms

Question

[fred.s.kremer]

Hello!

I'm beginning at RNA-Seq analysis and would like to do a differential expression analysis of some rice (O. sativa) datasets in different conditions. I looked at some protocols for DE analysis, and Trinity pipeline seems to be very straight-forward, although it was developed to be used in non-model organisms. Is it appropriate in my case?

[Ram]

AFAIK Trinity does assembly, not DE analysis. Have you looked at the actual DE tools that Trinity uses downstream (DESeq, edgeR, limma, voom etc?)

Answer 1

[h.mon]

I think it depends on the quality of the rice genome and annotation. If O. sativa has high quality genome and annotation, go for STAR + (edgeR | DESeq2), or (kallisto | Salmon) + sleuth. Transcriptome assemblies with Illumina reads are quite noisy, one gets many artefacts, and interpreting the results is more difficult. However, if the genome assembly is of poor quality, following a de novo pipeline may yield more complete results.

Assembling a transcriptome will demand a reasonable amount of computational resources, in particular, memory usage can get quite high if you have a lot of data. Trinity is an transcriptome assembler, but has many helper scripts to perform downstream analyses (including DE), and has an excellent wiki, with instructions covering from assembly to quality assessment to quantification and differential expression analysis.

My suggestion is to look at the recent rice literature and more or less follow the established analysis methods, updating programs and genome versions when appropriate.