Mapping coordinates from gff file to individual gene

Question

[GR]

Hi Guys,

I wanted to know if any utility exist to map the gff file coordinates to map on the individual genes. I mean, if gff file has coordinates for full genome, I want to convert these to coordinates for individual gene.

Thanks,

RT

Answer 1

[Michael]

That would be a a very simple tool: feature.coord - (gene.min - 1) gives you feature.coords relative to gene.min position.

Replace the landmark column with the gene id and you get a valid gff entry again.

[GR]

Hi Michael, Can you please tell the name of the tool? I can write my own python script to do this, just wanted to know if any tool exist before I write my own. Thanks.

Answer 2

[Alex Reynolds]

First, create a set of mappable genes (e.g., assuming human):

$ wget -qO- ftp://ftp.sanger.ac.uk/pub/gencode/Gencode_human/release_21/gencode.v21.annotation.gff3.gz \
    | gunzip --stdout - \
    | awk '$3=="gene"' - \
    | grep -f genes.txt - \
    | convert2bed -i gff - \
    > genes.bed

Then map your GFF file (my_annotations.gff) to these genes:

$ bedmap --echo --echo-map --delim '\t' <(convert2bed -i gff < my_annotations.gff) genes.bed > answer.bed

The answer will contain your GFF annotation, and all genes (and their coordinates) which overlap your annotation by one or more bases.

[GR]

Hi Alex, This is helpful but does not completely serve my purpose. For example if the gene is at position 1619..2809 in the genome. In my file I want to convert these coordinates in reference to individual gene which will be 1..1181 and similar for all the features of each gene. Do you know any tool/utility for this?

[Alex Reynolds]

If you need to do arithmetic on data in columnar text files, awk is a very useful tool. You can pipe data to it, as well as pipe data from it.

[GR]

Done with a python script! Thanks.