GC content normalization in discriminative de novo motif analysis

[Emi Ling]

Hello,

I am trying to perform discriminative de novo motif analysis using the command-line version of MEME-ChIP, and I am wondering whether MEME-ChIP automatically performs GC content normalization in such cases. My initial MEME-ChIP run identified several motifs with DREME that I am concerned are possibly due to biases in the local GC content (<30bp from the center of the supplied sequences). I have roughly 10 times as many control sequences as my primary ones, which hopefully allows the program to select an appropriately matched subset of control sequences. 

Thank you.

[noble]

MEME adjusts for the biases in letters and groups of letters using the background model
that you provide.  A 1-order model (made using fasta-get-markov) adjusts for dimer biases
(like GC).

DREME does not use a background model, and normalization depends on the control
dataset it is provide with.  MEME-ChIP uses fasta-shuffle-letters with -kmer 2, preserving
dimer frequencies.  You could try manually creating a -kmer 3 (or higher) set of shuffled
sequences, and rerunning DREME with them.  Refer to the "Program Information" section
of your MEME-ChIP output to see how you would do this.

Note that the MEME-ChIP web portal allows you to choose a higher-order background
model (for use by MEME) under "Universal Options".  That model is built from the input
sequences.  This does NOT currently affect the -kmer setting for the shuffled sequences
provided to DREME.