Confusion about one flag in sam format

[mimozz]

Hello,

I met one problem in the Tophat mapping. Does anyone know what the flag difference between 97 and 99 is?

The annotation of 99 is: read paired, read mapped in proper pair, mate reverse strand, first in pair

The annotation of 97 is: read paired, mate reverse strand, first in pair

Seems that the only difference is “read mapped in proper pair”. Is it due to the estimated insertion length? Can 97 be counted as a concordant pair?

Many thanks.

[swbarnes2]

"Proper pair" likely means that both reads are on the same chromosome, point in towards each other, and maybe that their distance is some appropriate number (I don't know how bowtie determines that, maybe the settings you give it in the command line)

So any of those things could be off, and give you a 97 flag.

So if the reads point away from each other, or are on different chromosomes, no, you should not count them as concordant.

[maubp]

The meaning of "read mapped in proper pair" is down to the aligner/assembler, and would consider the expected read pair orientation (typically --> <--, also <-- --> for mate pair, and --> --> or <-- <-- for native Roche 454 reads if not flipped to look like traditional paired reads) if mapped to the same reference, and the expected fragment/template length (i.e. separation of the reads, field TLEN aka ISIZE).