Problems About Music

Question

[yliubu]

Hi,

I am interested in the software of MuSiC and now trying to inplement it. When I use the calc-wig-covg function with the command of

Genome music bmr calc-wig-covg –wig-list=/home/share/input/wig_list.txt –output-dir=/home/share/output –reference-sequence=/home/share/input/allgenome_hg18.fa –roi-file=/home/share/input/roi.tsv

I got the following errors

Use of uninitialized value of $tot_covd in concatenation (.) or string at /usr/.../CalcWigCovg.pm line 229

And in all the output files, the covg of all samples in all genes are 0. I am not sure whether this error is due to my input file, so I attached some lines of each input file below:

Wig file:

TCGA-02-0003-01    /home/share/input/stddata_2012_08_25/GBM/Coverage/TCGA-02-0003-01.wig
TCGA-02-0033-01    /home/share/input/stddata_2012_08_25/GBM/Coverage/TCGA-02-0033-01.wig
...

reference file was downloaded from UCSC by

wget ftp://hgdownload.cse.ucsc.edu/goldenPath/hg18/chromosomes/*

and "cat" them all together:

>chr10
******
>chr11
****
...

ROI file:

chr1    87100715    87152695    SEPT15
chr2    109657664    109729072    SEPT10
chr2    241903395    241942114    SEPT2
chr4    78089918    78178792    SEPT11
chr4    164668498    164754226    MARCH1
chr5    10406827    10488491    MARCH6
chr5    126233449    126394145    MARCH3
...

Also, I found that ROI files must be in ordered, otherwise there are errors. But I am not quite sure based on which order, chromosome or start pos? For example, if I put "chr1 5 10 gene1" before "chr1 1 3 gene1", an error will show up said that the first start pos is bigger than the second one. So I am thinking if the two start pos are 10 and 5, which order should put them, based on the number value or number itself? And also, does all these orders obey chr 1-22, i.e., chr1 is all alway before chr2?

Can somebody explain the above problems? Thanks in advance.

[Neilfws]

Edited for readability. Please note how indenting code/output lines with 4 spaces helps.

Answer 1

[Cyriac Kandoth]

Make sure the roi-file and wig-list are properly tab-delimited and use unix-format newlines "\n". Can you also remove the chr-prefix from all your chromosome names. When calc-wig-covg parses wig files, it discards the chr-prefix, making it incompatible with your roi-file and reference-fasta that use chr-prefixes. All other tools in MuSiC are agnostic to chromosome naming conventions, so calc-wig-covg will be fixed in the next release.

If you installed MuSiC following the recommended instructions, then you also have a tool called joinx. Use joinx as follows to correctly sort your roi-file. The chromosomes can be sorted in any order, but the regions within each chromosome must be ordered by start loci.

joinx sort roi.tsv roi_sorted.tsv

Answer 2

[yliubu]

Thanks for your detailed information! That will be very helpful!

[Neilfws]

Please comment under the answer, rather than writing a new answer.