Find closest gene to chromosome location

Question

[D H]

Hello!

I started a gene enrichment analysis (I haven't done this before), and I have a dataset that contains the gene expression data. This data set has a column with the gene names.

However, there are some entries in that columns, which represent chromosome locations instead of gene names. I want to find the gene closest to these chromosomes locations.

I'm using R 2.15.2 (if that helps).

What are my options?

Thank you in advance!

Answer 1

[igor]

bedtools closest

Answer 2

[Nicolas Rosewick]

bedops closest-features: http://bedops.readthedocs.org/en/latest/content/reference/set-operations/closest-features.html

[D H]

I see that this utility takes as an input a BED file. I only have a csv file though.

Is it possible to convert csv to BED?

(I'm sorry for the questions but I'm very new to this)

[igor]

The most simple version of a BED file is three tab-separated columns (chr, start pos, end pos). You can use Excel to extract those three columns and save as tab-delimited text file.

[harold.smith.tarheel]

[Disparaging comment about Excel deleted]

[Alex Reynolds]

If you use Excel, be sure to clean it up. It can save tab-delimited text files, but with non-Linux line endings.

You can do the following fix, in the case of exporting from Excel on Mac:

$ tr '\r' '\n' < input.fromExcelForMac.txt | sort-bed - > input.fixed.bed

If you exported your data from Excel on Windows, apply this post-save fix:

$ tr -d '\r' < input.fromExcelForWindows.txt | sort-bed - > input.fixed.bed

Then run closest-features to query features of interest:

$ closest-features --closest input.fixed.bed features.bed > answer.bed