#1 Simulate longer Indels

[Colin Hercus]

Hi Heng Li,

I'd like to simulate reads where the mutations include longer insert/deletes and rather than write another program I thought it might be better to modify maq. The basic idea is to add a gap extend probability so that if a gap is to added we use rand() and Pextend to establish a length.
I've looked at maq code and think it could be done by using a 16-bit version of t_seq and using 4-bits to encode a mutation type, 0 delete, 1-6 for 1 to 6 bp encoded in last 12bits, 15 delete, 14 substitution. The remaining 12 bits would be 2-bits for the original character and 10 bits for up to 5 inserted bases.

I'm happy to do the changes and then submit back to you if you like.

Colin

[Colin Hercus]

Modified version of simulate.c

[Colin Hercus]

Logged In: YES
user_id=1613561
Originator: YES

Hi Heng Li,

I've gone ahead and made the changes. The code is attached if you'd like to use it.

Cheers, Colin
File Added: simulate.c

[lh3]

Logged In: YES
user_id=1602510
Originator: NO

Hi Colin,

Thank you very much for this. I am on the CSHL meeting, but I promise to have a look when I am back in UK.

Cheers,

Heng

[Colin Hercus]

Logged In: YES
user_id=1613561
Originator: YES

Hi Heng,

There's a bug in my code. I forgot to add X: into the optarg list for simulate. I did all my testing with default value for X.

I've also found that simustat needs to allow more slop in the alignment location for correct/incorrect decision.

Cheers, Colin

[Nobody/Anonymous]

Logged In: NO

Hi Heng,

Did you ever use this code?

Colin