MCAST Segmentation fault (core dumped)

[Jing]

Hi,

When I tried to run a simple MCAST command, I encountered the following error:

(10522) Reading sequences in blocks of 950 characters. Segmentation fault (core dumped)

The command I ran was:

mcast my_motifs.meme my.fasta

The file sizes are as follows:

• my_motifs.meme: 296 KB
• my.fasta: 245 KB

There is only one sequence in the my.fasta file.

I am using MEME Suite version 5.5.7, installed via Conda on Ubuntu 16.04.6 LTS (Xenial Xerus).

Do you have any thoughts on what might be causing this issue?

Thank you so much for your time.

Best,
Jing

[cegrant]

While we don’t currently support a Conda or Bioconda package, we do provide a MacPorts package and a Docker image. These are described in the installation guide. We are considering adding a Bioconda package at some point, have no definite plans at this point.

I can’t give you a definitive explanation for the segmentation fault you encountered, but I notice that your motif file is 296KB long. It’s actually larger than your sequence database! I suspect that MCAST is simply running out of memory. This may indicate a misunderstanding of MCAST’s function. I’m guessing that your motif file contains hundreds of different motifs and you are trying to find clusters involving any of them. In fact, MCAST will try to combine all of those motifs into a single model and try to find clusters containing as many of those motifs as possible. MCAST building the combined model for all those motifs would be what causes your system to run out of memory.

A more typical use for MCAST is to have previously identified a handful of motifs that are suspected to function together to regulate expression. MCAST then identifies clusters containing just those motifs. If you are just looking for candidate locations of transcription factor binding sites, FIMO would be the more appropriate tool. Be aware though, if you scan with hundreds of motifs your are introducing a significant multiple-testing issue into your results and you might need to apply a Bonferroni correction to the reported q-values.

Let us know if I’ve misunderstood your intent, or if you have further questions. It’s always helpful if you include copies of your input data along with your command line.

Charles