Sam FLAGS : how could a read be described as "read unmapped" but "read in reverse strand"?

Question

[johnnytam100]

I am so confused why an unmapped read could have a direction. Any explanation for that?

Thank you!

Answer 1

[glihm]

If the read is unmapped, the "strandness bit" has an other significance. From SAMTOOLS documentation:

Bit 0x10 indicates whether SEQ has been reverse complemented and QUAL reversed. When bit 0x4 is unset, this corresponds to the strand to which the segment has been mapped. When 0x4 is set, this indicates whether the unmapped read is stored in its original orientation as it came off the sequencing machine.

0x4 corresponds to the mapped/unmapped bit (If set, the read is considered as unmapped. If unset, the read is considered as mapped)

0x10 corresponds to the strandness bit

So, the translation of the documentation could be "If the reads is not mapped, the strandness bit indicates if the read (unmapped) is stored in itrs original orientation from the sequecing machine".

Which means, you can have an unmapped read with an orientation without any problem. You just have to give to it the good interpretation. ;)

[johnnytam100]

I see. Then what does "stored in its original orientation as it came off the sequencing machine" mean?

And how should I compare a read being described as "read unmapped" and a read being described as "read unmapped" and "read in reverse strand"? In short, one of them just give me one more description as "read in reverse strand", and practically they are the same because they are unmapped anyway?

[glihm]

Actually, I don't think you've to care about unmapped reads in general. Except that if the unmapped reads are your reads of interest (filtering data for instance).

practically they are the same because they are unmapped anyway?

Exactly. ;)

The only thing you have to keep in mind is that, for people interested in oriented sequencing, they are interested in knowing why some reads are unmapped and they want to keep the info of "which was the original orientation of the read during sequencing process". But for most of uses, unmapped reads are reads that are not meeting the parameters you have chosen for your alignments.

[johnnytam100]

Thank you very much! :)