sam flag 97 and 145

[hollandorange]

Hello,

After alignment, I got reads in the sam format with flag 97 and 145.
flag 97:
the read is paired in sequencing;
strand of the mate reverse;
the read is the first read in a pair;

flag 145:
the read is paired in sequencing;
strand of the query;
the read is the second read in a pair;

These reads are not mapped in a proper pair but both query and mate are mapped. how to explain this? could you give me an example?

Thanks
Yanju

[maubp]

They might be mapping to different contigs? What are the full SAM lines for these two reads?

[hollandorange]

The reads are looking like the following. I only have one reference sequence.

ENST000004134650_104_988_445_609_200# 97 ENSG00000141510 13261 37 36M = 14182 957 CGGTCAACCGTTTTGTAGAACAACTCCCGTCCCCTC 22222
2222222222222222222222222222222 XT:A:U NM:i:0 SM:i:37 AM:i:37 X0:i:1 X1:i:0 XM:i:0 XO:i:0 XG:i:0 MD:Z:36

ENST000004134650_104_988_445_609_200# 145 ENSG00000141510 14182 37 36M = 13261 -957 GCGGCCACATCCTCGACGACCACGTCCCCGGTGCCC 22222
2222222222222222222222222222222 XT:A:U NM:i:0 SM:i:37 AM:i:37 X0:i:1 X1:i:0 XM:i:0 XO:i:0 XG:i:0 MD:Z:36

[hollandorange]

anybody help?

[maubp]

First read, 97 == 1 + 32 + 64 == 0x1 + 0x20 + 0x40

0x0001 the read is paired in sequencing, no matter whether it is mapped in a pair
0x0020 strand of the mate (0 for forward; 1 for reverse strand)
0x0040 the read is the first read in a pair

Second read, 145 == 1 + 16 + 128 == 0x1 + 0x10 + 0x80

0x0001 the read is paired in sequencing, no matter whether it is mapped in a pair
0x0010 strand of the query (0 for forward; 1 for reverse strand)
0x0080 the read is the second read in a pair

So first read (flag 97) is on the forward strand (since it does not have 0x10 set), second read (flag 145) is on the reverse strand (since it does have 0x10 set).

[hollandorange]

I think it is the other way round:
the first read (97) is on the forward strand, since the strand of the mate is 1 (reverse);
the second read is on the reverse strand.

and in both cases (97,145)
0x0002 the read is mapped in a proper pair =0
0x0004 the query sequence itself is unmapped=0 (0 means it is mapped ???)
0x0008 the mate is unmapped=0 (0 means it is mapped ???)
This means both of them are mapped? but not proper?

[maubp]

I think it is the other way round:
the first read (97) is on the forward strand, since the strand of the mate is 1 (reverse);
the second read is on the reverse strand.

Yes - I had it right the first time, but flipped in back in a hurry before leaving the computer

[hollandorange]

I had a detail look at the samtool manu. I found, for these reads, the inferred insert size is very large like 957. but in my setting, the insertsize_mean is 200 and std is 20.

is it the reason that they are mapped, but their mapped position is too far away and then samtool report not proper mapped?

[nilshomer]

I had a detail look at the samtool manu. I found, for these reads, the inferred insert size is very large like 957. but in my setting, the insertsize_mean is 200 and std is 20.

is it the reason that they are mapped, but their mapped position is too far away and then samtool report not proper mapped?

The flags are set by the aligner, not samtools, so checkout the aligner's documentation. Check out the following to help you digest individual tags:
http://picard.sourceforge.net/explain-flags.html

Also, you can use the "-X" option in "samtools view" to convert the flag field to a string format.