Proper parameters for samtools view?

Question

[zhousun21]

Hi everyone,

I have a sam file created by minimap2: aln.sam

Using samtools view to convert it to a bam file works fine:

samtools view -S -b aln.sam > aln.bam

But adding the -f 4 parameter to retain the unmapped sequences does not work properly:

samtools view -S -b -f 4 aln.sam > aln.bam

It produces a very small file then the process ends. The small file can be sorted:

samtools sort aln.bam -o aln_sorted.bam

But when converting that to a bed file produces a file of size zero:

bedtools bamtobed -i aln_sorted.bam > out.bed

Can anyone see anything wrong with this command?

samtools view -S -b -f 4 aln.sam > aln.bam

Thanks for any advice.

Answer 1

[Istvan Albert]

If the reads don't align, there is nothing to convert to a BED file.

A BED file would contain the start and end coordinates of the alignments.

When you keep the unaligned reads, you basically keep alignments that are "not valid alignments".

[zhousun21]

I assumed that the -f 4 parameter would add the unaligned to the aligned, and keep both.

No? It keeps only the unaligned?

[ATpoint]

The -f is a filter, it means either keep (-f) or remove (-F) those reads with a given flag. Your -f 4 means to keep only the unmapped reads in the BAM. if you have few of them then small BAM size is expected. Unmapped reads have no coordinates, and BED is a coordinate-based format, hence you cannot convert unmapped reads to BED.

[zhousun21]

So, not adding any -f or -F flags will keep everything, but the unmapped reads will not appear in the bedfile regardless because there are no associated coordinates.

Is that right?

[ATpoint]

Yep :)

[zhousun21]

Ah! Thanks very much for clarifying. That explains it.