How to deal with AB SOLiD paired end reads?

Question

[biock]

Hello, I'm new in analyzing RNA-seq data and I need to process several sampled sequenced by AB SOLiD system in paired end mode (SRR1175538). These fastq files are not similar to Illumina fastq,
1. Reads length are not the same in forward and reverse fastq, read length of forward is 75nt while the reverse is 35nt
2. Many reads in the reverse fastq file are NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN

Should I trim these reads before alignment and what tool should I use? And what aligner should I use for reads mapping? Can I just use STAR or hisat2 in their PE mode?

Thanks a lot.

[colindaven]

Actually, the read 2 data in SOLiD were pretty much absolute rubbish. Their chemistry never really worked for the R2.

I would completely ignore them and use the R1 only.

Answer 1

[h.mon]

Don't trim the reads, SOLiD works best when you map to a reference genome the whole reads. You have to use a colorspace mapper, probably Subread will do a good job at it - STAR and HISAT2 are basespace only mappers. You may try to map the reads as paired-end and just the R1 as single-end, and see which works best, but as colindaven said above, working with just R1 may be better.

[colindaven]

Yes, good point. SOLiD reads have internal "frameshifts" which make them utterly useless for de novo assembly. Mapping to a good genome is absolutely essential.