segmentation fault with covariates file

[girdhar3]

Hey I am trying to run rasqual on chip-seq data and whenever i use my covariates file I get segmentation fault error. it works file without covariates file

tabix $VCF chr5:150219491-150290291 | rasqual -y $H3K4me3_counts_bin_file -k $H3K4me3_offsets_bin_file -s 150226191 150233313 150283713 -e 150226801 150233614 150285429 -j 30657 -l 922 -m 150 -n 30 -z -f "ACC_Plus_K4me3" -x $H3K4me3_covariates_bin_file -t
Segmentation fault

Also
for some features I am getting -nans in one of the column as an output

for example

tabix $VCF chr3:2102865-3016784 | rasqual -y $H3K4me3_counts_bin_file -k $H3K4me3_offsets_bin_file -s 2102865 2137383 2260304 2290182 2303466 2487799 2493511 2507834 2547112 2552350 2620889 2698716 2700440 2769101 2806883 2848422 2864056 2878623 2889979 3003245 3016088 -e 2103648 2144544 2260915 2290948 2304207 2488705 2494483 2508317 2548954 2557195 2621291 2699616 2700722 2769757 2807188 2849692 2865110 2879473 2890935 3003732 3016784 -j 15305 -l 13115 -m 1000 -n 14 -z -f "ACC_Plus_K4me3" -t

I get

ACC_Plus_K4me3 . chr3 2804401 A G 0.321429 3.623884 1.000000 -0.0001814579 2.3395617589 0.672682 0.010000 0.500000 1.679557 -0.102658 0 3333 4 2 2804401 -49.655566 0 -nan 0.994438

Can you help with this error?

[dg13]

Thanks for using the software. Can you make (very small) input files (the output of the tabix command you're running, the counts_bin, and offsets files) and upload them to dropbox / google drive so we can replicate your issue? Thanks

[natsuhiko]

Hi, it seems your covariate file contains binary values whose length is not multiple of 30 (sample size). Could you check the number of columns and rows of your original covariates file?

In addition, the feature start and end arguments (-s and -e) must by given by comma separated values instead of space separated.This is because the R-square value becomes NaN.