How to retrieve nearest genes in Ciona genome?

Question

[zhenyisong]

I have a batch of genome positions of conserved DNA regions from Ciona genome. I want to know if there exists a program to pin down the nearest gene. Thanks.

The genome position information is from VISTA-Point and looks like the following which is annotated Conserved Non-coding Sequence (CNS).

>C. intestinalis v.2.0 chr_01p:14451-14569 (+)
AAGTTTTCAAAGTGGTGAAGAAATGAAGCACCCTCGTATTTATTAAGTTTAGCACAAGTC
TGTTCCACACAACTTGTCATAGGAGGTGGCCCGCAACTGAAGACGCCAAACTTTTCAAC
>Ciona savignyi Sep. 2005 reftig_173:198261-198379 (-)
AAGTTTTCAAAATGATGCAGGAAGCTTGGACCTTCATACTTGTTGAGTTTGCTGCATGTT
TTGTCCACCTCATTGGTCATTGGAGGAGGACCACAACTGAAGACTCCGAACTTGTCAAC
=  length = 119bp, identity = 71.4%, type = exon

Answer 1

[Alex Reynolds]

If your regions and genes are in sorted BED format, or if they can be converted and sorted, you can use BEDOPS closest-features to find the nearest upstream and downstream genes to each region. BEDOPS also includes conversion and sorting utilities, if that functionality is needed to prepare your data.

[zhenyisong]

BEDOPS is a beautiful program. It seems that I have to prepare all required BED (or starch format) files for genes and target DNA region first and then I can use closest-features to locate the nearest gene. But I want to know who else already done this for us to parse Ciona genome? So I can use them immediately in one command line. Anyway, thanks for the advice.

Answer 2

[Hugues]

The gene list for Ciona intestinalis is available at http://www.ncbi.nlm.nih.gov/gene/?term=ciona%5BOrganism%5D

Click Sent To: > File > Format: > Tabular(text) to save it on the disk (gene_result.txt).

This file is not in BED format so you'll need to convert it first.

I wrote a python script to help you, copy-paste and save it as genelist_to_bed.py

$ python genelist_to_bed.py

Usage: genelist_to_bed.py input_file.txt output_file.bed

Here is the code:

import csv
import os
import sys

if len(sys.argv) < 3:
    sys.exit('Usage: %s input_file.txt output_file.bed' % sys.argv[0])

if not os.path.exists(sys.argv[1]):
    sys.exit('ERROR: Gene list %s was not found!' % sys.argv[1])

fieldnames = ('chromosome',
              'start_position_on_the_genomic_accession',
              'end_position_on_the_genomic_accession',
              'GeneID',)

fi = open(sys.argv[1], 'rb')
fo = open(sys.argv[2], 'wt')

try:
    reader = csv.DictReader(fi, delimiter='\t')
    writer = csv.DictWriter(fo, fieldnames=fieldnames, delimiter='\t')
    headers = dict( (n,n) for n in fieldnames )
    fo.write('# ')    
    writer.writeheader()
    for row in reader:
        writer.writerow({ headers[fieldnames[0]]:"chr"+row[fieldnames[0]],
                          headers[fieldnames[1]]:row[fieldnames[1]],
                          headers[fieldnames[2]]:row[fieldnames[2]],
                          headers[fieldnames[3]]:row[fieldnames[3]],
                          })
finally:
    fi.close()
    fo.close()

Then you can use the tool suggested by Alex Reynolds.

Hope it helps!

[zhenyisong]

They are different gene assembles, NCBI vs. VISTA data (my raw data from VISTA which contains the genome fragment position information). I once thought it was a small task, but then find it is a project. I have to re-map VISTA coordinates (I don not know which assemble they use) to NCBI. You script let me know what the minimum information the BED file is. Thanks.

[Hugues]

Can you edit your question so that it contains the necessary details?

Also, where does the genome positions of conserved DNA regions comes from? Is it C. Intestinalis v.2.0 - Ciona savignyi Sep. 2005 provided in VISTA-point?

FYI the minimum information for the BED file is described in the BEDOPS help.

[zhenyisong]

Your are right. The position information is from VISTA-Point. I can handle it now. The step is to use BLAST to re-map the CNSs to the NCBI Ciona genome assemble as I have no other idea of how to translate VISTA-Point coordinates to NCBI's. Anyway, Thank you and Alex.

[Hugues]

You could create a new question: how to map VISTA coordinates to NCBI (ciona)?

and cross-ref it here. I don't know how to tackle that though, but I'd be interested to learn about what you find out.