Hello,

I have about 100 bacterial species. I used "reutils" package to download their fasta files. I am now trying to download their gff files. I have the name and NC_ ids for these.

I looked online and see that that there is package "genomes" that can read the file into R, but you need to provide the complete link to it . Eg: file<-"ftp://ftp.ncbi.nih.gov/genomes/Bacteria/Yersinia_pestis_CO92_uid57621/NC_003132.gff"

If the link had been something like "ftp://ftp.ncbi.nih.gov/genomes/Bacteria/NC_017105.gff", it would have been easy to programmatically create the link  in R and download the file. But that is not the case, since there is a species name and another number in it.

1) Is there an "retuils" type package where I can give the NC_ id and download the gff file ?

2) Is there a way to create this link programmatically for each bacteria ?

Any insights ? Thanks

You can get the directory listing using curl and ftp

library(RCurl)
curl <- getCurlHandle()

url <- "ftp://ftp.ncbi.nih.gov/genomes/Bacteria/"
xx <- getURL(url=url, ftplistonly=TRUE, curl=curl)
entries <- strsplit(xx, "\n")[[1]]

parse these for the species names

ncbi_taxa <- sub("^([[:alpha:]]+_[[:alpha:]]+)_.*", "\\1", entries)

and match them up with the information you have

keep <- match(your_taxa, ncbi_taxa)
gff_urls <- paste0(url, entries[keep], "/", your_NC[!is.na(keep)])

you could then use download.file() or RCurl and this suggestion to download each gff. Import the downloaded files, or the url's to the files directly, with rtracklayer::import().