Samtools view error:

Question

[jaqx008]

Hello all,

I am having an error which is new to me and I couldn't really find any useful tips.

I am trying to convert a sam file generated with bowtie (one) into a bam file using samtools view -b. The command prints the following code and a 27byte bam output. see error bellow. Do anyone know what might be the issue or how I can rectify this?

[E::sam_hrecs_error] Header line does not have a two character key at line 246:
"@A00881:421:HM7W5DRXX:1:2101:1118:1016 16  NC_005109.4 73902262    255 23M GTCAACATCAGTCTGATAAGCTA IIIIIIIIIIIIIIIIIIIIIII XA:i:0  MD:Z:23 NM:i:0  XM:i:2"
samtools view: failed to add PG line to the header ^Z

Thank you

[GenoMax]

See if this applies: https://github.com/samtools/samtools/issues/1237

[Istvan Albert]

The error may be fixable by editing the SAM file header.

[jaqx008]

@ Istvan Albert I saw that I can edit the header with samtools header but not sure what the new header should read (what to include).

[jaqx008]

Thank you. The link recommended adding --no-PG to the command. This did not really help. it produced the error

error: node035: task 0: Exited with exit code 1

[Istvan Albert]

what is the output of

cat data.sam | head -100 | grep ^@

that would tell you which headers seem malformed.

Answer 1

[arrmino]

I guess the problem is that the name of reads have the @ as start and samtools interprets that as a header line. Probably this happens if you run for example bowtie2 using a FASTA instead of a FASTQ file, i.e. for FASTA the @ in read names does not get removed. Use sed to edit the file like: sed 's/\@A00881/A00881/g' yoursamfile > newsamfile

Answer 2

[jilguero888]

It seems that samtools is becoming stricter about proper format of sam files about tags in header lines. I had a similar error to the one reported here just because I did not set properly the SM tag in the @RG header line when calling bowtie2.

wrong: --rg-id idxx --rg text right: --rg-id idxx --rg SM:text

It worked to me.