Given loci coordinates, is it possible to identify gene name from hg19.2bit file?

Question

[akylzhanov.r]

Hi everyone.

I have a bed file with N specified loci. Is it possible to look up what gene corresponds to a given loci in the bed file?

If yes, how can I do the request to the ucsc server not downloading whole genome?

Many thanks.

[GenoMax]

How to identify which genes fall within genomic locations
N Closest Genes To A Given Location

Answer 1

[anp375]

I assume you want all genes in a given locus. First, download the latest release of the appropriate GTF. You don't need the binary genome. This is what you can do using bash shell:

• Download and install bedtools. Then load it.
• Use "grep" to extract the genes only. "grep -e $'\tgene\t' /path/to/hg19.gtf > /path/to/hg19_genes.gtf"
• Convert the bed file to the correct format: chr\tstart\tstop\tname with regex if necessary. Output this to Genes.tsv.
• Then do bedtools intersect. "bedtools intersect -a /path/to/genes.tsv -b /path/to/hg19_genes.gtf -wao -bed | tr -d '"' | tr ';' '\t' > intersect.bed
• Cut out the necessary columns. Include strandedness, which I don't think I included. GTFs only use forward coordinates and you'll need to take the reverse compliment of those specified sequences. "cut -f 1-4,6,13,14,18 intersect.bed | tr ' ' '\t' > present_genes.tsv"
• For other objectives, you may need to concatenate the regions in the files to get contiguous regions, which I'm unfortunately not allowed to help you with.
• You will most likely have to customize these commands. 'gene' which is in the same column as 'CDS', 'start_codon', 'exon', etc., may be called something else. You should also read the documentation for bedtools to find out what's going on. Good luck.

[anp375]

I'm pretty sure you can do this on Galaxy as well, using the same commands, but I don't know how. I find it a pain to use.

[akylzhanov.r]

Thanks, for you answer, anp375. I have hg19.2bit file. How do i convert between them in first place?

[anp375]

You can't convert that. A GTF doesn't have the actual sequences in it. It's just an annotation file with gene names, coordinates, etc. You can either the latest release for the annotations:

ftp://ftp.ensembl.org/pub/release-75/gtf/homo_sapiens/

or the earliest:

ftp://ftp.ensembl.org/pub/release-55/gtf/homo_sapiens/

Note that these are from ensembl and the instruction were meant for that. You can get something from UCSC over here: http://genome.ucsc.edu/cgi-bin/hgTables?command=start where you select the hg19 assembly, genes and gene predictions, region genome, track whatever names you want, and select GTF in the output format.