bedops overlap output

Question

[bioguy24]

I am just curious why the output of the below bedops overlap command results some single gene per line and on others 2 or 3 or 4. Thank you :).

bedmap --echo --skip-unmapped --delim '\t' --echo-map-id-uniq unix_sort_4column_xgen_targets.bed sort_gene_exon.genecode.bed > answer.bed

chr1    979192    979413    +    AGRN:exon.10
chr1    979478    979647    +    AGRN:exon.11
chr1    979703    979829    +    AGRN:exon.12
chr1    980530    980667    +    AGRN:exon.13
chr1    980728    980913    +    AGRN:exon.1;AGRN:exon.14
chr1    981102    981266    +    AGRN:exon.15;AGRN:exon.2
chr1    981333    981478    +    AGRN:exon.16;AGRN:exon.3
chr1    981529    981655    +    AGRN:exon.17;AGRN:exon.4

Answer 1

[Alex Reynolds]

For some elements of the file unix_sort_4column_xgen_targets.bed, only one exon from the file sort_gene_exon.genecode.bed overlaps that element, and so one exon ID is reported.

For other elements of the file unix_sort_4column_xgen_targets.bed, two exons from the file sort_gene_exon.genecode.bed overlap that element, and so two exon IDs are reported.

Because you use the --skip-unmapped option, elements from unix_sort_4column_xgen_targets.bed that have zero overlaps with elements from sort_gene_exon.genecode.bed are not reported in the output file. You are skipping unmapped reference elements.

[bioguy24]

So for

chr1    980728    980913    +    AGRN:exon.1;AGRN:exon.14

exon 1 and exon 14 are pulled down by the capture probe? Thank you :).

[Alex Reynolds]

If the regions of exons 1 and 14 overlap that reference element, then yes. If you're not sure about your input data, do spot checks of specific elements. Also make sure that you are using BEDOPS sort-bed to sort your BED files, because 1) Unix sort is slower and 2) most uses of sort I see on this site do not sort on the third column, while sort-bed does.

[bioguy24]

Thank you :).