Problem with BED file in methylKit

[bbm]

Hello all!

I am using methylkit to analyze my methylation data in honey bees.
I have converted a reference genome annotation file (gff3 file) to a bed file, then I run

gene.obj=readTranscriptFeatures(system.file("extdata", "amel_OGSv3.2_sorted.bed",
package = "methylKit"))

The msg was that:
Calculating intron coordinates...
Error in convertBed2Introns(bed) : this is not a valid bed file.


My file is like this:
Group1.1 377 436 . . - amel_OGSv3.1 CDS 2 Parent=GB42164-RA
Group1.1 377 436 . . - amel_OGSv3.1 exon . Parent=GB42164-RA
Group1.1 377 79932 GB42164 0.786254 - amel_OGSv3.1 gene . ID=GB42164
Group1.1 377 79932 GB42164-RA 0.786254 - amel_OGSv3.1 mRNA . ID=GB42164-RA;Parent=GB42164
Group1.1 650 969 . . - amel_OGSv3.1 CDS 0 Parent=GB42164-RA

Should I only keep the first 3 required columns, and test it again?
Appreciate your thoughts.

[Michael.Ante]

Hi bbm,

I think the problem is not the name field, but the missing information on exon-intron structure (since the function "convertBed2Introns" called the error).
Parsing a GFF3 file into something well usable is always hard. I'd go for converting your gff3-file into a gtf file, which is then easier to convert to a bed-file with exon-intron information.

Cheers,

Michael

[bbm]

Thank you for your thoughts. I have downloaded a bed file of honey bee from UCSC and tested in R, but it's still not working. This bed file has all 12 columns.
The only difference is my bed file has "Group1" as the chromosome # in the first column, but the example file has "chr20" in the first column.


Error in convertBed2Introns(bed) : this is not a valid bed file

Hi bbm,

I think the problem is not the name field, but the missing information on exon-intron structure (since the function "convertBed2Introns" called the error).
Parsing a GFF3 file into something well usable is always hard. I'd go for converting your gff3-file into a gtf file, which is then easier to convert to a bed-file with exon-intron information.

Cheers,

Michael