Dear Subread team

I have some questions about Subread with -B option.

Since the multi-mapping reads are defined as reads with >1 equally-best mapping locations, what happens if I use the -B <n> option to limit the number of mapping sites reported for Subread (or -nBestLocations in RSubread)? Are <n> mappings reported randomly?

How would you advice to use the -B option if my data as about 40-50% repeat mappings, and i wish to count all equally-best mapping locations for a read, in order to quantify repeats properly.

Also, how is this affected by -Q or -H options?

Thanks

Vivek

Dear Vivek,

Equally-best mapping locations are ordered by their hash values in the lookup table and locations with smaller hash values will be reported first. Number of mismatched bases is used in finding equally-best mapping locations. -Q and -H options are no longer used in the latest version (1.5.0).

You might try different -B values to see how many more alignments you will get when you increase your -B value. Once you get to the number where there is no or little increase on the reported number of alignments, you should get the threshold you need for mapping.

Best,

Wei