Can Soap2 Be Used For Rna-Seq Analysis

Question

[Wligtenberg]

Hi,

I want to re-analyse results of RNA-seq data. Previously, they have been analysed using SOAP2 and an old reference genome. When I contacted the author of snpEff (because I want to check the SNP effects later on) he mentioned that he was unsure if SOAP2 could be used to analyse RNA-seq data, because it can't map across exon boundaries. So, can SOAP2 be used for RNA-seq mapping or not? Or can it be used, but does another tool (like TopHat) a really better job?

Thanks in advance.

Answer 1

[Malachi Griffith]

If you want to align RNA-seq reads with an aligner designed for DNA mapping you can create a custom reference genome that contains the chromosomes plus a database of exon-exon junctions. This deals with the issue of mapping reads across exon-exon junctions, but it limits your detection to those junctions that you define in your database. Several groups have used this approach to align RNA-seq reads with BWA and other aligners that are not 'splice-aware'. If your reads are long (>75 bp), the simplest solution is probably just to use TopHat to align your reads against a standard reference genome. It will definitely do a better job than simply using SOAP2 against the standard reference genome. In addition to TopHat there are many other 'splice-aware' aligners designed with RNA-seq reads in mind. These include: TopHat, SpliceMap, MapSplice, hmmSplicer, Supersplat, SOAPsplice, etc.

Of course there are also many other older splice aware aligners that would produce useful results but are too slow to be practical when aligning the number of reads typical in an RNA-seq experiment. These include: BLAT, Exonerate, Spidey, Splign, etc. They might still be useful in the context of performing an evaluation of the next-gen splice-aware aligners with a test data set.

[Wligtenberg]

So, in short. Using SOAP2 with a standard reference genome is not a good idea, right?

[Malachi Griffith]

Right, not a good idea. Not for RNA-seq in any case.

Answer 2

[Istvan Albert]

In general one cannot take a tool designed for DNA mapping and directly use it for transcriptome data unless the data has certain characteristics or the questions that need to be answered are very simple.

You will need to find a tool that can handle the job. Note that the tasks may be quite challenging as shown for example in Transcript assembly and quantification by RNA-Seq... (Nature Biotechnology, 2010)