Fpertille Fpertille - 3 months ago 14
R Question

How can I extract sequences from a FASTA file for each of the intervals defined in a BED file using R?

How can I extract sequences from a FASTA file for each of the intervals defined in a BED file using R?
The reference genome used is "Gallus gallus" that can be obtained by:

source("http://bioconductor.org/biocLite.R")
biocLite("BSgenome.Ggallus.UCSC.galGal4")
library(BSgenome.Ggallus.UCSC.galGal4)


My data file is a result of gRanges package

library("GenomicRanges")

> olaps
GRanges object with 2141 ranges and 0 metadata columns:
seqnames ranges strand
<Rle> <IRanges> <Rle>
[1] chr14 [ 1665929, 1673673] *
[2] chr14 [ 2587465, 2595209] *
[3] chr14 [ 8143785, 8151529] *
[4] chr14 [ 9779705, 9787449] *
[5] chr14 [10281129, 10288873] *
... ... ... ...
[2137] chr24 [3280553, 3288297] *
[2138] chr24 [3330889, 3338633] *
[2139] chr24 [3005641, 3015321] *
[2140] chr24 [3319273, 3327017] *
[2141] chr24 [5549545, 5557289] *
-------
seqinfo: 31 sequences from an unspecified genome; no seqlengths


That I can transform in data.table

olaps<- as.data.table(olaps)


Example to be used:

olaps<-"seqnames start end width strand
chr1 1665929 1673673 7745 *
chr1 2587465 2595209 7745 *
chr1 8143785 8151529 7745 *
chr2 9779705 9787449 7745 *
chr2 10281129 10288873 7745 *"
olaps<-read.table(text=olaps,header=T)


Expected outcome:
something like this (fasta format):

>SEQUENCE_1
ACTGACTAGCATCGCAT...
>SEQUENCE_2
ACGTAGAGAGGGACATA...
>SEQUENCE_3...


I have tried to use this package unsuccessful until now:

source("http://bioconductor.org/biocLite.R")
biocLite("rtracklayer")

Answer

This, should solve your trick:

First:

seq = BSgenome::getSeq(BSgenome.Ggallus.UCSC.galGal4, olaps)

to add names to the sequences:

names(seq) = paste0("SEQUENCE_", seq_along(seq)) 

To generate a ".fasta" from your sequences:

Biostrings::writeXStringSet(seq, "my.fasta")

More details were provided before:

https://support.bioconductor.org/p/77913/#77986