How to remove Read Name (1st column) extension from a BAM file?

Question

[IrK]

Hello everyone,

i have paired-end data, after alignment I have noticed that R1 and R2 have different read names in BAM file, for example:

SRR1032070.122660125.1
SRR1032070.122660125.2

so read R1 has extension .1 and read R2 has extension .2. This causes a problem when I try to convert BAM to BED file with bedtools bamtobed -bedpe -i. So the only solution I can think of is to remove these extensions. Could anyone please advice on the tool, I would not like to convert data back to SAM, as it is extremely large!!!

Thank you

[GenoMax]

I assume you got this data from SRA? You should have used -F|--origfmt Defline contains only original sequence name option to avoid getting these kind of read names.

As for adding /1 /2 to read names you could use reformat.sh from BBMap suite with the addslash=t or addcolon=t options.

[IrK]

ohhh ok, so you mean when I convert SRA to FATSQ with fastq-dump -F use F option

Thank you

Answer 1

[Pierre Lindenbaum]

samtools view -h in.bam | sed '/^[^@]/s/^\(.*\)\.[12]\t/\1\t/' | samtools view -Sb -o out.bam -

[IrK]

thank you so much, Pierre.

I modified the sed part [12] to [.1.2] and it works. So the final command, which works for me is:

samtools view -h in.bam| sed '/^[^@]/s/^\(.*\)\.[.1.2]\t/\1\t/' | samtools view -Sb -o out.bam

May I clarify for the future reference, how paired-end reads can be annotated with extensions 1 and 2? What software was used for this?