maq merge limit

[Layla]

Hi,

Is there a limit to the number of map files that can be merged using
./maq mapmerge? If so, is there a way to overcome this for short paired end reads?

Cheers

L

[nilshomer]

Hi,

Is there a limit to the number of map files that can be merged using
./maq mapmerge? If so, is there a way to overcome this for short paired end reads?

Cheers

L

I would assume the limit is the maximum number of arguments generally a command line tool can accept (argv & argc stuff), which is in the thousands. Anyhow, if you have millions of .map files, I would ask why? If you truly have millions of .map files you could just do the merge step of merge sort whereby you group .map files into pairs of .map files and merge each pair, rinse and repeat. Again, how many .map files do you have?

Nils

[Layla]

Thank you for your reply
I have 10 patients each generate 15 .map files

I split them into pairs of 1 million (total 2 million reads) to speed up maq alignment.

L

[nilshomer]

Thank you for your reply
I have 10 patients each generate 15 .map files

I split them into pairs of 1 million (total 2 million reads) to speed up maq alignment.

L

15 .map files no problem.

[Layla]

Great, but would you know of any way to speed maq up? Even chunks of 2 million take a few days as I am running the .map command simultaneously. If not, not a problem.

Cheers
L

[der_eiskern]

i've merged 41 map files 'lickedy split'

[der_eiskern]

Layla what's your machine set up? i can map 40 sets of 1 million reads in 5.5 hours using a single processor on a machine with 16 Gb RAM. are you running it simultaneously on multiple processors?

[henry]

Hi,

Is there a limit to the number of map files that can be merged using
./maq mapmerge? If so, is there a way to overcome this for short paired end reads?

Cheers

L

it's no problem to merge 15 map files for maq.
BWA adopted burrows-wheeler transform algorithm and many tricks to accelerate processing speed. BWA is also written by the author who coded maq. You can try BWA. it should be faster.