Is sorting by read name processed in the case of single reads with multiple mappings?

[martina]

Hi Yang,

I'm wondering how sorting works on single reads with multi-mapping: is sorting by read name in this case processed as in the case with paired-end reads? 

I ask you this because I did some tests activating (1) and deactivating (2) the sorting by-readName in a position-sorted input bam file:

1) $ featureCount -M -a <annot file path> -o <result files path> <pos sorted reads>.bam  

2) $ featureCount -M --donotsort -a <annot file path> -o <result files path> <pos sorted reads>.bam

and I get ~same execution time in both cases while I was expecting to get an higher exe time in case 2), where the fast sorting routine should be deactivated.

Is the --donotsort arg only used combined with pair-end reads?

Thank you in advance!

Cheers,

Martina