Question about paired-end alignment

[Pia]

Hello everyone, I have a question about how STAR aligns paired-end reads.

I noticed that in the BAM output file the field RNEXT/MRNM always points to the same chromosome of the read that I am looking at. With other aligners (e.g. bwa) there are usually some cases where Read1 and Read2 are aligned to different chromosomes.

Does STAR always align the reads adjacently (i.e. not independently) ? Or are read pairs that are not on the same chromosome filtered out? Or does STAR simply write the same chromosome into the RNEXT/MRNM field even if that might not be how it was actually aligned (i.e. this field has no information)?

Thanks for your help!

[Alexander Dobin]

Hi Pia,

by default, STAR does not output such "chimeric" alignments. You would need to switch it on with --chim* options.

Cheers

Alex