RNA-Seq Bowtie2 alignment shifted left 1 position.

[seanb89]

I'm exploring some 100bp SE RNA-seq data using Bowtie2 and a short reference sequence given to me.

When I use the --very-sensitive-local argument in Bowtie2 I get successful mappings, but with a huge number of mismatches. When viewed in IGB/IGV at the nucleotide level, it seems that the mappings are shifted 1 position to the left of the reference.

I've not encountered this before with Bowtie2, and was wondering if this was a feature of the local mapping algorithm, or do I have some other fundamental misunderstanding of the algorithm?

[Brian Bushnell]

Cross-posted here.