Split reference genome into callable regions by splitting on NNNN stretches

Question

[William]

I am looking for a tool or piece of code that can take in a reference genome fasta file and output a bed file with X (equally sized) callable regions. The output can also be anything similar to the content of such a bed file.

These callable regions should be bordered upstream and downstream by any of these:

• The start of the chromosome
• The end of the chromosome
• a stretch of 250bp unknown (N) nucleotides (or e.g. 1000bp N)

This to get regions that can be variant called in parallel, without the risk of sequencing reads and variants going over the borders of the callable regions.

Answer 1

[Pierre Lindenbaum]

• use picard ScatterIntervalsByNs to extract the non-poly N regions.
• convert this interval list to bed with awk.
• run bedtools makewindows to generate your regions.