Is it normal a substantial reduction of BAM files after ordering ?

Question

[Antonio R. Franco]

I am getting BAM files with Bismark mapping with bowtie2. This gives you an unordered BAM file Thus, I am sorting the BAM file with the command

samtools sort -O bam -T tmp_ -o Name_sorted.bam Unordered.bam

The unordered bam file can have a 2,6Gb size, but the ordered one drops down to 1,2Gb. The same happens with all the bam files I am ordering. They are reduced to a half.

Is that normal? Is that caused by a different compression level ?

In any case, makes me feel sort of uncomfortable..

[ATpoint]

Use samtools flagstat to check the number of reads. If they are the same, you're fine and it is due to compression. How did you create the unsorted bam file?

[Antonio R. Franco]

I mentioned that. I am mapping BS-Seq sequences with Bismark

[grant.hovhannisyan]

Yes, it is normal. Sorting can save a lot of space.

Answer 1

[andrew.j.skelton73]

Sorting helps to give a better compression ratio. There's an old discussion about it here, or a tip of the day here

edit: I should also mention that you'll see a much better compression ratio with RNA seq than with DNA, as RNA seq should have a lot of repeated sequences.

Answer 2

[Antonio R. Franco]

You are right. They have the same number of reads