Make a BedGraph file

Question

[kirillkirilenko]

I have a simple BED file displays repetitive elements localization (this file was obtained using RepeatMasker).

Here are the first few rows::

2L  840 882 Simple_repeat
2L  2763    2796    Simple_repeat
2L  3409    3498    DNA/Helitron
2L  5405    5432    Simple_repeat
2L  8844    9132    LINE/unknown

Now, I would like to create a BedGraph file that displays the coverage of repetitive elements per selected bin. Let's assume the bin size is 3000. The desired file should look like this::

2L      1       3001    0.025
2L      3001    6001    0.0387

where 0.025 is calculated as ((882-840)+(2796-2763))/3000.

I can write a Python code to accomplish this, but I believe there might be ready-made tools available for this task. Could you assist me? Thank you in advance.

Answer 1

[bhumm]

Bedtools window piped into genomecov?

https://bedtools.readthedocs.io/en/latest/content/tools/window.html

https://bedtools.readthedocs.io/en/latest/content/tools/genomecov.html

Or perhaps genomecov paired with awk? I'm thinking something like:

genomeCoverageBed -d -bg | awk 'begin{sum=0;} {sum=sum+1;if(NR%3000==0){print sum;sum=0;}}'

Disclaimer: I have not tested either of these strategies

Answer 2

[rfran010]

I would suggest bedtools makewindows to generate your 3,000bp bins then bedtools coverage to generate fraction of repeats over your windows. Maybe:

bedtools makewindows -w 3000 -g chrom_sizes > windows.bed
bedtools coverage -a windows.bed -b repeats.bed > coverage.bed
cut -f1-3,7  coverage.bed > coverage.bg

the cut command is because bedtools coverage outputs 4 extra columns, with the forth being "The fraction of bases in A that had non-zero coverage." If you use this command or otherwise have a minimal bed, columns 1-3 & 7 will produce your bedgraph. Otherwise, you will need to adjust column number of the fraction.

[kirillkirilenko]

It works, thank you!