PeakSeq peakselect segmentation fault

[crackerjack.tej]

Hi,

I'm trying to use PeakSeq v1.1 to call peaks on Illumina reads. I used the preprocess option to process both my Input and Experiment sam files. I downloaded the mappability file of hg19 (my reads were mapped to chr1-22, X, Y, and M) from Gernstein Lab's website. But when I run the peak_select option, I get a segmentation fault in the middle. When I rerun, it stops at exactly the same place. I have 6 different experiment files, and they all behave the same way, although the place at which they give the error is different from one another. I thought it may have something to do with the input, so I tried using the -validate_fragments. But I don't understand what am I to do with its output. I don't know what is giving this error, and any help is really appreciated. Here is my config.dat file:

Experiment_id N2UO_Peaks
chromosome_list_file hg19_chr_list
Enrichment_fragment_length 200
target_FDR 0.05
N_Simulations 10
Minimum_interpeak_distance 200
Mappability_map_file mappability_hg19.txt
ChIP_Seq_reads_data_dirs SO_2365_N2UO_Bound
Input_reads_data_dirs SO_2365_INPUT_pooled
Simulation_seed 1234567
Background_model Simulated
max_Qvalue 0.05

Let me know if any other info is needed.

Thanks a lot,

TEJ

EDIT: Figured out the reason. The chrM reads were somehow creating an issue. When I reran without chrM in the list, there were no issues

[maciosz]

Hi,

I have the same problem, except the "figured out" part. I get segmentation fault in the same exact moment every time. It also helps to remove some of the chromosomes, so the problem isn't in mappabilty file or program itself, but since I've got to remove most of the chromosomes i wouldn't exactly call it a solution. Do you have any ideas what can cause such a behaviour and how to avoid it? Other programs don't seem to have problem with my input files, just PeakSeq.

[crackerjack.tej]

I think the problem comes when you use the wrong mappability file (a file made from wrong set of chromosomes), though I am not sure. The reason I figured out chrM was the prob was because every file stopped when PeakSeq was processing chrM. Can you look into the output of PeakSeq and see where exactly is that stopping? Also, did you download the mappability file, or created one for yourself?

[maciosz]

I created one myself. I have several experiments to analyse and in each of them different chromosomes cause problems - In one of them, for example, program stops for every chromosome except 3. i 4., and for the other it manages to go through all of them except Mt. So I don't suppose it's caused by a mappabilty file