(Modern, mid-2016) RNA-seq software pipeline

Question

[jp]

I am trying to follow the RNA-seq protocol described in (this excellent paper):

Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks (2012) http://www.nature.com/nprot/journal/v7/n3/full/nprot.2012.016.html

This was published three/four years ago, so not that much in terms of scientific timeline - but I noticed the tools described in the pipeline have not been updated in a while or have been replaced by other software. More specifically:

• TopHat has entered a low maintenance, low support stage as it is now largely superseded by HISAT2 (ref: https://ccb.jhu.edu/software/tophat/index.shtml)
• The github (https://github.com/cole-trapnell-lab/cufflinks) repo for cufflinks hasn't been recently updated (in about a year)
• Samtools is now at version 1.3.1, but this doesn't work with the current version of cufflinks (0.1.19 works though)

I know there are other software components which can replace some of the above (e.g. Star aligner instead of bowtie/tophat). I understand that the standard answer is "depends on what you want to do" (differential gene expression and variant analysis in my case - from human samples).

My question to the community is - what is a modern bioinformatics software pipeline to carry out RNAseq analysis?

Thanks for your suggestions or comments you may have,

[GenoMax]

• Use a splice-aware aligner if you are not using bacterial samples (STAR is popular but needs good bit of RAM, BBMap is simple to use).
• Count using featureCounts (part of subread package, much faster than htseq-count, will automatically sort bam files). Make sure SAM tags are in v.1.3 format (most counting packages don't understand SAM v.1.4, the current format)
• DE analysis with DESeq2 (or edgeR) or a comparable package.

Much the same advice is here Suggestions for RNA seq pipeline

[fanli.gcb]

This post also has more information, particularly re: pseudoalignment approaches

[vakul.mohanty]

You might want to take a look at kallisto it's quantifies gene expression without alignment, making it very fast. https://pachterlab.github.io/kallisto/

Answer 1

[ivivek_ngs]

I would would suggest also to take a look at Salmon followed with DESeq2 /edgeR/ voom for Differential analysis , be it at transcript level or gene-level. You can also take a look at this link which has summarization of different RNA-Seq tools which have evolved over the years.

Answer 2

[amit.sinhaa]

An updated version of the TopHat-Cufflinks pipeline (Trapnell et al 2012), with all the updated tools has recently been published as: Pertea et al (2016) Transcript-level expression analysis of RNA-seq experiments with HISAT, StringTie and Ballgown. www.nature.com/nprot/journal/v11/n9/full/nprot.2016.095.html Hope you find it useful

[User 59]

This feels more appropriate for the OP as they have specifically said they want to do downstream variant analysis which implies at some point you're going to want to align the data, rather than rely on pseudoalignment approaches (awesome though they are)

Answer 3

[Carlo Yague]

There is the RNA-seq workflow at the gene level from bioconductor (2015). Its main focus is the differential expression analysis stage but it also briefly covers the read mapping and summarization.

Answer 4

[Ming Tommy Tang]

I keep a note for RNA-seq here https://github.com/crazyhottommy/RNA-seq-analysis#normalization-quantification-and-differential-expression