Statistics on BED files

Question

[abedkurdi10]

Hello everybody, I have two BED files. I would like to know which one of them has closer intervals. I already calculated standard deviation for the distances between consecutive intervals. First, I calculated the distances between consecutive intervals using bedtools spacing. since I have different numbers of records in both files, I don't know how to compare standard deviation to make the real comparison.

Does anybody knows how to make the comparison? Is there any other ways?

I appreciate your help, Thank you.

Answer 1

[Alex Reynolds]

I think one way is to compute and compare the coefficients of variation for your two files, which are unitless or dimensionless.

Start with set A:

$ cut -f1-4 A.bed | sort-bed - > A.bed4
$ closest-features --closest --dist --no-overlaps --delim '\t' A.bed4 A.bed4 > A.dist.bed5
$ awk '{a+=$5; n[NR]=$5;} END {m=a/NR; s=0; for(i=1;i<=NR;i++) {s+=((n[i]-m)*(n[i]-m));} sd=sqrt(s/(NR-1)); cv=sd/m; print cv;}' A.dist.bed5 > A.cv.txt

Repeat for set B:

$ cut -f1-4 B.bed | sort-bed - > B.bed4
$ closest-features --closest --dist --no-overlaps --delim '\t' B.bed4 B.bed4 > B.dist.bed5
$ awk '{a+=$5; n[NR]=$5;} END {m=a/NR; s=0; for(i=1;i<=NR;i++) {s+=((n[i]-m)*(n[i]-m));} sd=sqrt(s/(NR-1)); cv=sd/m; print cv;}' B.dist.bed5 > B.cv.txt

Then compare A.cv.txt and B.cv.txt to get the relative variability.

[abedkurdi10]

I will give it a try. Thank you.