[dpryan]

That's what I've read but upon looking on the actual BAM outputs, this is not the case (anymore, probably).

The flags for single-end experiment are limited to 0,16,256 and 272 (which I'd imagine would be + and - reads aligned "as is" and + and - reads aligned with a junction).

chadn737 was talking about the MAPQ column (the 5th column), not the flag (2nd column)

[apredeus]

chadn737 was talking about the MAPQ column (the 5th column), not the flag (2nd column)

God, I should read more carefully. Thank you

[apredeus]

If you are using 1.4.1 then the mapping scores should all follow this formula:

255 = unique mapping
3 = maps to 2 locations in the target
2 = maps to 3 locations
1 = maps to 4-9 locations
0 = maps to 10 or more locations.

You can use this info then to quickly count reads for each category.

This actually does not seem to hold upon checking (or at least not anymore). In Tophat 1.4.1, there is no mapping quality of 2. I've looked at a large number of BAM files and it seems that the qualities are as follows:

255 = unique mapping
3 = maps to 2 locations in the target
1 = maps to 3-4 locations
0 = maps to 5 or more locations (up to the number defined in "--max-multihits").

cheers