How to extract upstream and downstream genes of a list of locations from gtf file?

Question

[aboozar.soorni]

I have a list of locations in a bed format

chr2    55159107    55160004
chr3    40280597    40282177
chr4    74682484    74683574
chr4    76795449    76796456
chr6    10250838    10251741
chr6    20435795    20436466
chr6    31169498    31170294

I am interested to identify genes that are 1000bp (1K) upstream and downstream of each of those location from a gtf file, which belongs to a non model plant.

gtf file:

SpoScf_00032    maker   exon    12116   12419   .   +   .   gene_id  transcript_id "Spo06120";
SpoScf_00032    maker   exon    14070   17062   .   +   .   gene_id  transcript_id "Spo06120";
SpoScf_00032    maker   exon    17626   17899   .   +   .   gene_id  transcript_id "Spo06120";
chr2    maker   CDS 15262965    15263150    .   +   0   gene_id  transcript_id "Spo26212";
chr2    maker   CDS 15264530    15264667    .   +   0   gene_id  transcript_id "Spo26212";
chr2    maker   CDS 15265433    15265885    .   +   0   gene_id  transcript_id "Spo26212";

bedtools window, intersect, closest doesn't answer my question because they look for overlaps.

[Bastien Hervé]

You can create a bed_downstream file with chr, start-1000, start and an other bed_upstream file with chr, end, end+1000.

Then you can run bedtools intersect once on bed_downstream then on bed_upstream.

Answer 1

[Pierre Lindenbaum]

get the upstream / downstream regions of the bed

awk '{X=1000;B=(int($2)-1)-X;if(B<0) B=0;printf("%s\t%d\t%s\n",$1,B,$2);printf("%s\t%s\t%d\n",$1,$3,int($3)+X);}' the.bed

sort and use bedtools intersect

[aboozar.soorni]

Thanks so much,

Do I need to apply for specific options in bedtools intersect?

[Pierre Lindenbaum]

yes. read the manual.

[aboozar.soorni]

I used your awk script to get up and down regions of my bed files, which looks like:

bed file:

SpoScf_00500    226344  227695
SpoScf_00562    236367  239437

Result:

SpoScf_00500    0   226344
SpoScf_00500    227695  228695
SpoScf_00562    0   236367
SpoScf_00562    239437  240437

This is correct? for the first line, 225344 should not be instead of 0?!

[Pierre Lindenbaum]

fixed . changed P->B