My pair-end RNA-seq data is stranded (fr-firststrand, dUTP library). I tried to align the reads onto the genome using:
HISAT2:
hisat2 -p 32 --dta -x genome_index -1 1P.fq.gz -2 2P.fq.gz -S map.sam
and STAR:
STAR --runThreadN 12 --readFilesCommand gunzip -c --readFilesIn 1P.fq.gz 2P.fq.gz --genomeDir star_index --outSJfilterReads Unique
However, when I check the results from the two aligners on IGV, it seems that STAR aligned the reads on the wrong strand:
The upper panel shows read junctions from STAR and the lower panel is from HISAT2.
Is it the case that STAR aligned the reads on the wrong strand? How do I fix this? Thanks
Looks like you did not add correct --rna-strandedness RF option for hisat2 alignment. Default option is to consider data "unstranded".
See: Which strand specific option need to be used with HISAT2?
Hi, yes I realised that, but in this case, hisat2 (the lower panel) did align the reads to the correct strand - see gene model below. But STAR aligned the reads on the opposite strand.
dUTP method sequences the complementary strand: HISAT2 rna-strandness option
I see! It makes sense now, thank you.
I ran HISAT2 with --rna-strandness RF again, but the alignment looks the same. In fact, it looks more messy than the previous run without the strandeness option - a few reads appear to map to the reverse strand.
I am so confused, in this case, HISAT2 (with and without the strandness option) always show opposite strandness compared with STAR, how this can be potentially explained?
Log in to answer this question.
