RNA-seq alignment software

[rakesh12]

Hi All,
We are working on RNA-seq data. Our aim is to perform expression quantitative trait loci analysis for these samples. I am trying to find the most optimal method for alignment with very good sensitivity and specificity for RNA-seq data. I am trying to compare four different software for alignment namely MAQ, BWA, Bowtie and Bowtie-2. Can any one of you please suggest which alignment software might be more suitable for RNA-seq? Also, it would be very helpful if you can also provide the comparison in terms of advantages and disadvantages for these software for aligning RNA-seq data.
Thanks for your kind help,
Rakesh

[kopi-o]

I wouldn't use any of them, unless you map against a transcriptome reference. (If you map against the genome, you will miss splice junctions.) With a transcriptome reference, you won't be able to see novel transcripts or splice variants.

TopHat and many other programs (MapSplice, SpliceMap, RUM, Subread...) can do spliced mapping to the genome.

If you have to use one of your 4 programs, I would choose BWA over MAQ (because they have the same author and BWA is newer) and Bowtie2 over Bowtie (for the same reason).

[epi]

I generally agree with the person posted above. However, I will factor in read length and reference genome. depending on those, you can still go ahead with your approach in a fraction of time it will take you to do splice alignment

[rakesh12]

Hi Kopi,
Thank you so much for your detailed response. I completely agree with you. I am thinking about using Bowtie2 to align RNA-seq data against transciptome reference. Then use RUM for spliced mapping. So, I can quantify gene expression as well isoform expression. Which method should I use for bias correction in terms of total numbers of reads, gene length or normalization?
Also, what are the key differences between BWA and MAQ in terms of sensitivity and specificity apart for been a hash table based vs. Burrows-Wheeler transform algorithms.
Thanks again for your help,
Rakesh