MAST backgrounds

[Hai Li]

When I use fasta-get-markov , I set the order 8.But when I use this 8-order background file to run MAST,MAST cannot work.

• My command for fasta-get-markov is:
  fasta-get-markov -m  8  /home/meme/aa.fa  >  /home/meme/aa_bg

• My command for mast is :
  mast  -nostatus  -remcorr  -ev 10.0  -bfile /home/meme/aa_bg /home/meme/motifs.meme  /home/meme/aa.fa
  
  
  
• the error is:
  mm_resize(new_array->items,num_items,ATYPE) failed!

But when I set the order of background to 5,MAST works correctly.What can I do if I set the order to 8 .

The another problem is that when I use '-comp' is MAST, the error is:

mast: pssm-distr.c:28: calc_pssm_pdf: Assertion `get_array_length(prob) >= alen' failed.

In my opinion, '-comp' allows me to use each sequences' frequencies to get the p-value and E-value.

Thank you!

[CharlesEGrant]

MEME is the only program in the MEME Suite that supports background models higher than 0th order, and for MEME we don't recommend using more than a 3rd order model. MAST and the other programs only support 0th order background models.

The issue the '-comp' option turns out to be a bug in MAST introduced when we added support for arbitrary alphabets. Our apologies for the inconvenience. We'll set that this gets fixed in the next release of the MEME Suite. In the meantime I've attached a patch file that you can use to fix the bug. To apply the patch, copy the patch file to the root of the MEME Suite distribution, and run the commands:

patch -p1 -i mast.patch
make
make install

Note that MAST will always report p-values and E-values for the top, matching, sequences. The -comp option corrects the reported  p-values and E-values for the observed composition of the sequence. Note that this option does not affect the computation of the scores and p-values for the individual motif matches, only p-value for the overall sequence match. 

[Hai Li]

Thank you so much!