Bowtie: ?'s about h_sapiens_37_asm index/outputs

[cutcopy11]

Hello,

1)
I have a questions about bowtie's "h_sapiens_37_asm" index.
For the human genome, the concise reference sequences are numbered 0 to 24. I
assume these numbers represent chromosomes. Is 0 chromosome 1? What about X,
Y, and the M?

bowtie -z -f h_sapiens_37asm --concise reads/seq_human.fa human_map

2)
I am also confused about the localization of the 5' end of the read.

According to the dafault output description, the fourth column shows the
"0-based offset into the forward reference strand where leftmost character of
the alignment occurs."

If you don't use the concise option ...

if a read is mapped 5' to 3' from position 1000 to 1035,
the fourth column should be 1000, right?

If a read is mapped 5' to 3' from position 1035 to 1000,
will the fourth column be 1035 or 1000?
I think this means the fourth column will be 1000.
If this is the case, is there an "option" so that I can make it 1035 instead of 1000?

Thanks,
Clayton

[Ben Langmead]

Hi Clayton,

Sorry - I saw your email and meant to reply immediately but it's probably better to reply here anyway...

1)
I have a questions about bowtie's "h_sapiens_37_asm" index.
For the human genome, the concise reference sequences are numbered 0 to 24. I
assume these numbers represent chromosomes. Is 0 chromosome 1? What about X,
Y, and the M?

bowtie -z -f h_sapiens_37asm --concise reads/seq_human.fa human_map

It's 0=chr1, 1=chr2, ..., 21=chr22, 22=X, 23=Y, 24=MT. Not wanting anyone to have to contend with these numbers is one of the reasons I'm likely to deprecate --concise soon (I've already deprecated -b/--bin).

2)
I am also confused about the localization of the 5' end of the read.

According to the dafault output description, the fourth column shows the
"0-based offset into the forward reference strand where leftmost character of
the alignment occurs."

If you don't use the concise option ...

if a read is mapped 5' to 3' from position 1000 to 1035,
the fourth column should be 1000, right?

If a read is mapped 5' to 3' from position 1035 to 1000,
will the fourth column be 1035 or 1000?
I think this means the fourth column will be 1000.
If this is the case, is there an "option" so that I can make it 1035 instead of 1000?

Thanks,
Clayton

You're right that it will be 1000 in both cases. There's no option to invert that for reads on the reverse strand, but it's easy to convert Bowtie's output to your desired output with a little postprocessing. An example:

$ ./bowtie --quiet -c e_coli GGTGCGTTACTGTGGCACTTCGNCCTGATGTGGAT,GCATTTTTTTCGCCAGCCAGGCTTTCGCTTTGGGT
0 - gi|110640213|ref|NC_008253.1| 4769855 ATCCACATCAGGNCGAAGTGCCACAGTAACGCACC IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 0 22:G>N
1 + gi|110640213|ref|NC_008253.1| 2365447 GCATTTTTTTCGCCAGCCAGGCTTTCGCTTTGGGT IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 0
$ ./bowtie --quiet -c e_coli GGTGCGTTACTGTGGCACTTCGNCCTGATGTGGAT,GCATTTTTTTCGCCAGCCAGGCTTTCGCTTTGGGT | awk -v OFS='\t' '{if($2 == "-") {$4 += (length($5)-1)} ; print $0}'
0 - gi|110640213|ref|NC_008253.1| 4769889 ATCCACATCAGGNCGAAGTGCCACAGTAACGCACC IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 0 22:G>N
1 + gi|110640213|ref|NC_008253.1| 2365447 GCATTTTTTTCGCCAGCCAGGCTTTCGCTTTGGGT IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 0

Ben

[cutcopy11]

Thanks Ben