How To Estimate K And Lambda For Local Alignment'S Significance?

Question

[Tianyang Li]

Hi,

I'm wondering how to estimate K and lambda for local alignment's significance for a specific scoring matrix and gap penalty. I know there's a paper (The statistical parameters for local alignment score distributions) about it, but I hope that there's already a tool I can use directly. Does anyone know such a tool? Thanks.

[Tianyang Li]

I asked Dr. Altschul and he suggested that I use by Dr. Spouge's group. The tool they provided (alp1.3) couldn't be compiled with g++ 4.4, so I fixed a few things (added a few headers) and it can now be compiled using g++. The code is here.

[Tianyang Li]

I asked Dr. Altschul and he suggested that I use by Dr. Spouge's group for gapped alignment (ALP on http://www.ncbi.nlm.nih.gov/CBBresearch/Spouge/html_ncbi/html/index/software.html). The tool they provided (alp1.3) couldn't be compiled with g++ 4.4, so I fixed a few things (added a few headers) and it can now be compiled using g++. The that works with g++ is here.

[Tianyang Li]

I asked Dr. Altschul and he suggested that I use the tool by Dr. Spouge's group for gapped alignment (ALP on http://www.ncbi.nlm.nih.gov/CBBresearch/Spouge/html_ncbi/html/index/software.html). The tool they provided (alp1.3) couldn't be compiled with g++ 4.4, so I fixed a few things (added a few headers) and it can now be compiled using g++. The that works with g++ is here.

Answer 1

[Casey Bergman]

I had this same problem a number of years back. I emailed Stephen Altschul who sent me some C code to calculate K and lambda, which worked fine as I recall. Perhaps you can try this same thing and also see if Altschul would be willing to post this code more publicly.