R session aborted / fatal error when running read_homer from a list of motifs about universalmotif HOT 5 CLOSED

I understand -- no worries about not having the option to update. Sorry I couldn't figure out what was wrong in the end. Glad to hear read_matrix() didn't crash for you (which puzzles me slightly, but alas).

Should you ever update to the newest version and still encounter this kind of bug, feel free to reopen this issue and I'd be happy to try and investigate again.

from universalmotif.

Hi,

Thanks for reporting this. My guess would be an error happening somewhere during the process of generating the universalmotif-class motif within read_homer() (that portion is done using C++ code, which is the most likely to cause a sudden crash out of R). Unfortunately it's not something I can more accurately diagnose without a bit more information.

First, if you can, please update to Bioconductor v3.14 (this would also require updating to R v4.1). This is because I can only issue bug fixes to the current Bioconductor version, and there is also a small chance whatever this bug is has been fixed since then.

Second, if it's not too much trouble, please try and narrow down which motif(s) are causing the problem -- i.e. in your loop just also include a line to print which file is being read or some such. If you're able to narrow this down to a single motif which is causing trouble, please just paste it here and I'm certain I could easily fix the problem.

Hopefully this isn't too onerous, I'd be glad to figure out what's causing this and fix it.

(Edit: Unless I'm misunderstanding and actually the crash is happening some time after having read all the motifs?)

from universalmotif.

Hi!
I was able to narrow down which files from 2 different lists are causing the problem. Here are the 2 motif files:
File 1:

>VNNGGATTADNN	bcd(Homeobox)/Embryo-Bcd-ChIP-Seq(GSE86966)/Homer	6.298101	-6.564805	0	T:195.0(17.52%),B:6222.5(14.26%),P:1e-2
0.343	0.229	0.252	0.176
0.302	0.189	0.292	0.217
0.217	0.298	0.279	0.206
0.205	0.080	0.629	0.086
0.001	0.001	0.997	0.001
0.696	0.302	0.001	0.001
0.001	0.001	0.001	0.997
0.001	0.068	0.001	0.930
0.816	0.001	0.001	0.182
0.313	0.120	0.326	0.240
0.227	0.281	0.298	0.195
0.239	0.246	0.246	0.269

File 2:

>NNRCCTAACT	BOS1(MYB)/col-BOS1-DAP-Seq(GSE60143)/Homer	5.846626	-6.199610	0	T:228.0(15.21%),B:5785.3(12.65%),P:1e-2
0.195	0.331	0.209	0.265
0.234	0.204	0.240	0.322
0.447	0.047	0.482	0.023
0.101	0.641	0.148	0.110
0.027	0.901	0.032	0.040
0.246	0.100	0.210	0.445
0.888	0.025	0.041	0.046
0.714	0.232	0.027	0.027
0.118	0.574	0.081	0.227
0.257	0.239	0.054	0.450

Thank you so much for your time and help!!

Anne-Sophie

from universalmotif.

Thanks.

Unfortunately no matter what I do I can't get it to crash trying to read the motifs, so I'm afraid it could be a setup-specific bug. I've tried both on macOS and Linux. Just to confirm, did you try this again with the newest version of universalmotif? (I see you are on macOS yourself. While I was able to test the same version as you are using on Linux, this is not something I can do easily on macOS, so I could only test using more recent versions of the package.)

Also, just in case: does trying to read in the motifs using read_matrix() instead also crash? E.g.:

read_matrix("motif.homer", header = ">", positions = "rows")

from universalmotif.

Hi,
it looks like the version I'm using of universalmotif is 1.6.4, I'm assuming that's the latest version available for the Bioconductor and R versions I'm using as there does not seem to be newer versions available when I try to update it. I can't update Bioconductor/R for now (I will definitely do as soon as my papers are published).

I tried using read_matrix, and I do not get any error, R does not crash, and it successfully generates a formal class 'universalmotif' object, even for the motifs that were causing R to crash. I will use this function instead.... Not sure what caused the problem. Thank you so much again for your help, I really appreciate!!

Anne-Sophie

from universalmotif.