What is the fundamental difference between STAR and Bowtie(2). Specifically, what is the difference in their final output (regardless of run-time differences, speed, memory usage etc.). Both seem to work on the same kind of input *.fastq files without discriminating what they are associated with and generate an output version of the *.fastq files. Would my downstream analysis be incorrect if I use Bowtie2 where STAR was supposed to be used, and vice versa? Thank you.
What is the fundamental difference between STAR and Bowtie(2).
STAR for mapping spliced (i.e. with introns) short RNA-seq reads against a genome. Bowtie2 for mapping short reads without splicing.
Would my downstream analysis be incorrect if I use Bowtie2 where STAR was supposed to be used (?)
Yes. If you use bowtie2 to map spliced short reads against a genome, STAR-based downstream analysis is likely to be wrong.
... and vice versa?
If you map genomic reads with STAR, the result will be suboptimal. Use the right tool for the right task.
The seeds are clustered together by proximity to a selected set of ‘anchor’ seeds in the 2nd phase of the STAR algorithm, but this •clustering/stitching/scoring process• is not present in Bowtie2.