BEDTools: Unexpected file format. Please use tab-delimited BED, GFF, or VCF.

id0 · 08-19-2013, 07:45 AM

I am using BEDTools and getting the following error:

Code:

Unexpected file format.  Please use tab-delimited BED, GFF, or VCF. Perhaps you have non-integer starts or ends at line 1?

I had a GFF file and figured maybe there is indeed a problem, so I converted it to a 3-column BED file just to be safe. The BED file did not work either.

I searched around for this error and it's usually caused by non-unix newlines or non-tab separators. My file does not have either of those issues.

Here are the first few lines of the BED file:

Code:

chr1	3214482	3216968
chr1	3421702	3421901
chr1	3670552	3671498
chr1	4290846	4293012
chr1	4343507	4350091

Any idea what may be causing this error?

GenoMax · 08-19-2013, 07:49 AM

Check the bed file using the following command to make sure there are no unexpected characters.

Code:

$ cat -e filename.bed | more

or

Code:

$ cat -A filename.bed | more

The tab characters are displayed as "^I" (at least on CentOS).

id0 · 08-19-2013, 07:51 AM

Quote:

Originally Posted by GenoMax

Check the bed file using the following command to make sure there are no unexpected characters and the file is indeeded tab limited.

Code:

$ cat -e filename.bed | more

I already did that. No unexpected characters.

GenoMax · 08-19-2013, 07:57 AM

Are you getting the error even with a small subset like the one you pasted above?

id0 · 08-19-2013, 08:05 AM

Used different files and small subsets and still getting the same problem.

Supposedly, the problem is at line 1, but there is really nothing wrong with it as far as I can tell.

GenoMax · 08-19-2013, 08:11 AM

Not sure which exact bedtools program you need but perhaps bedops (http://code.google.com/p/bedops/) could provide an alternative.

id0 · 08-19-2013, 08:27 AM

I am using coverageBed. I just wanted to quickly get coverage at specific regions.

I have a valid BED file. Why is BEDTools disagreeing?

GenoMax · 08-19-2013, 08:29 AM

Have you used this copy of bedtools successfully before?

id0 · 08-19-2013, 08:33 AM

No. Unfortunately I just got BEDTools.

It compiled and it seems to be running okay for the most part. I assume if there was a problem with BEDTools, it wouldn't launch in the first place or produce some other lower-level error first.

Heisman · 08-19-2013, 09:42 AM

Create a file with vim and then type the following (with tabs):

chr1 500 1000
chr1 700 1200

and try to use "mergeBed" on it and see if that works.

id0 · 08-19-2013, 10:41 AM

Quote:

Originally Posted by Heisman

Create a file with vim and then type the following (with tabs):

chr1 500 1000
chr1 700 1200

and try to use "mergeBed" on it and see if that works.

mergeBed works on that example. It also works on other BED files.

coverageBed does not work on any of the files.

This is very confusing.

wadachi · 01-15-2014, 12:53 AM

How about checking "new line characters"?

if you make your .bed files on Windows, new line characters should be set \r\n.
whereas Unix/Linux uses only \r, so the excess \n may make BEDTools confused.

to justify this, choose line end character of Unix/Linux in the selection pulldown in your text editor's "save as ..." dialogue. (in my case, gedit in Ubuntu).

hope this works!

inesdesantiago · 06-30-2014, 06:03 AM

You could also use GenomeIntervals2BED.py script available within the SeqGI framework.
Have a look: http://seqgi.sourceforge.net/Genomeintervals2bed.html

jshaik · 04-07-2015, 02:54 PM

use dos2unix command in linux to convert your bed file to a file that is acceptable by bedtools. This happens as the windows and linux have different new line characters

Armin_P · 06-22-2015, 05:15 AM

I was facing the same problem. One of my files was a bam file and the other gff3/bed (I do not know if you have a bam file). Anyways, I used 'bedtools coverage' with the -abam option instead of -a and provided my bam file there. -b was then the gff3 or .bed (I tried both) and it worked!

swbarnes2 · 06-22-2015, 03:16 PM

In the future, you really need to include the exact command line you are using.

I'm pretty sure the problem is this: BED format is valid with only three columns of data, but I think coverageBed needs four columns of data; I think it wants a column for direction.

Try adding a fourth column of all pluses.

aditisk · 02-22-2016, 06:43 AM

@swbarnes2, does the 4th column need to be the direction or can it be gene names as well ? I have a .BED file that has some annotation information in the 4th column. Would that still be okay ?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Is there a tool that converts TXT, BED, GFF format to VCF?	LauraSmith	Bioinformatics	4	03-22-2017 02:41 AM
Updated How to convert .txt file to .bed .GFF or .BAR file format,	forevermark4	Bioinformatics	2	06-30-2014 06:02 AM
htseq--SAM line does not contain at least 11 tab-delimited fields	FLYINGDOLPHIN	Bioinformatics	0	06-03-2013 12:04 PM
creating tab-delimited output from unix	tonybert	Bioinformatics	2	01-27-2013 07:37 AM
Tab Delimited File Editors? (GFF to GTF)	DrD2009	Bioinformatics	16	11-30-2012 05:52 AM

08-19-2013, 07:45 AM	#1
id0 Senior Member Location: USA Join Date: Sep 2012 Posts: 130	BEDTools: Unexpected file format. Please use tab-delimited BED, GFF, or VCF. I am using BEDTools and getting the following error: Code: Unexpected file format. Please use tab-delimited BED, GFF, or VCF. Perhaps you have non-integer starts or ends at line 1? I had a GFF file and figured maybe there is indeed a problem, so I converted it to a 3-column BED file just to be safe. The BED file did not work either. I searched around for this error and it's usually caused by non-unix newlines or non-tab separators. My file does not have either of those issues. Here are the first few lines of the BED file: Code: chr1 3214482 3216968 chr1 3421702 3421901 chr1 3670552 3671498 chr1 4290846 4293012 chr1 4343507 4350091 Any idea what may be causing this error?

08-19-2013, 07:49 AM	#2
GenoMax Senior Member Location: East Coast USA Join Date: Feb 2008 Posts: 7,080	Check the bed file using the following command to make sure there are no unexpected characters. Code: $ cat -e filename.bed \| more or Code: $ cat -A filename.bed \| more The tab characters are displayed as "^I" (at least on CentOS). Last edited by GenoMax; 08-19-2013 at 07:52 AM.

08-19-2013, 07:57 AM	#4
GenoMax Senior Member Location: East Coast USA Join Date: Feb 2008 Posts: 7,080	Are you getting the error even with a small subset like the one you pasted above?

08-19-2013, 08:05 AM	#5
id0 Senior Member Location: USA Join Date: Sep 2012 Posts: 130	Used different files and small subsets and still getting the same problem. Supposedly, the problem is at line 1, but there is really nothing wrong with it as far as I can tell.

08-19-2013, 08:11 AM	#6
GenoMax Senior Member Location: East Coast USA Join Date: Feb 2008 Posts: 7,080	Not sure which exact bedtools program you need but perhaps bedops (http://code.google.com/p/bedops/) could provide an alternative.

08-19-2013, 08:27 AM	#7
id0 Senior Member Location: USA Join Date: Sep 2012 Posts: 130	I am using coverageBed. I just wanted to quickly get coverage at specific regions. I have a valid BED file. Why is BEDTools disagreeing?

08-19-2013, 08:29 AM	#8
GenoMax Senior Member Location: East Coast USA Join Date: Feb 2008 Posts: 7,080	Have you used this copy of bedtools successfully before?

08-19-2013, 08:33 AM	#9
id0 Senior Member Location: USA Join Date: Sep 2012 Posts: 130	No. Unfortunately I just got BEDTools. It compiled and it seems to be running okay for the most part. I assume if there was a problem with BEDTools, it wouldn't launch in the first place or produce some other lower-level error first.

08-19-2013, 09:42 AM	#10
Heisman Senior Member Location: St. Louis Join Date: Dec 2010 Posts: 535	Create a file with vim and then type the following (with tabs): chr1 500 1000 chr1 700 1200 and try to use "mergeBed" on it and see if that works.

01-15-2014, 12:53 AM	#12
wadachi Junior Member Location: japan Join Date: Jan 2014 Posts: 1	How about checking "new line characters"? if you make your .bed files on Windows, new line characters should be set \r\n. whereas Unix/Linux uses only \r, so the excess \n may make BEDTools confused. to justify this, choose line end character of Unix/Linux in the selection pulldown in your text editor's "save as ..." dialogue. (in my case, gedit in Ubuntu). hope this works!

06-30-2014, 06:03 AM	#13
inesdesantiago Member Location: LONDON, UNITED KINGDOM Join Date: Jan 2009 Posts: 44	You could also use GenomeIntervals2BED.py script available within the SeqGI framework. Have a look: http://seqgi.sourceforge.net/Genomeintervals2bed.html

04-07-2015, 02:54 PM	#14
jshaik Junior Member Location: st. louis Join Date: Jun 2011 Posts: 6	use dos2unix command in linux to convert your bed file to a file that is acceptable by bedtools. This happens as the windows and linux have different new line characters

06-22-2015, 05:15 AM	#15
Armin_P Junior Member Location: Austria Join Date: Sep 2014 Posts: 4	I was facing the same problem. One of my files was a bam file and the other gff3/bed (I do not know if you have a bam file). Anyways, I used 'bedtools coverage' with the -abam option instead of -a and provided my bam file there. -b was then the gff3 or .bed (I tried both) and it worked!

06-22-2015, 03:16 PM	#16
swbarnes2 Senior Member Location: San Diego Join Date: May 2008 Posts: 912	In the future, you really need to include the exact command line you are using. I'm pretty sure the problem is this: BED format is valid with only three columns of data, but I think coverageBed needs four columns of data; I think it wants a column for direction. Try adding a fourth column of all pluses.

02-22-2016, 06:43 AM	#17
aditisk Member Location: Ann Arbor Join Date: Aug 2015 Posts: 12	@swbarnes2, does the 4th column need to be the direction or can it be gene names as well ? I have a .BED file that has some annotation information in the 4th column. Would that still be okay ?