Sorting an already sorted file with samtools reduced its size

Question

[shhsvri]

Hi all

I used the

STAR ... --outSAMtype BAM   SortedByCoordinate ...

option while mapping my paired end .fastq files.

I then sorted it using samtools

samtools sort -O bam -o STAR_output.sorted.bam STAR_output.bam

The size of the bam file shrank by 25% from 8G to 6G after sorting with samtools. their default sort is sorts by coordinate (position).

I assume that the sorted bam file from STAR is the same as the samtools sorted bam file. though the latter is more compressed. Is that true?

[finswimmer]

There are several compression level for bam. Maybe samtools and start using different as default.

But why do you use samtools sort if star give you already a coordinate sorted bam file?

Answer 1

[h.mon]

I assume you are correct, as the documentation of both tools explaining the default values are in line with your hypothesis.

For samtools sort, the default value is -1, which means it uses the default for zlib, which is usually 6. You can change the compression level with -l.

For STAR, the default compression level is 1, which is the lowest compression. You can change the compression level with --outBAMcompression.