Working with the genome of non-model organism in R Bioconductor

GenomicFeatures and BSgenome are two widely used packages for genomic data analysis in R Bioconductor. They have variety of functions to work with genome features. Unfortunately, there are only a limited species you can work with BSgenome. But, you can create Genome Features and BSgenome for your organism of choice if it is not packaged into BSgenome by default.

First let’s create a taxonomic database using GenomicFeatures package. I am using Orf virus as an example here .

First, download GFF3 file of the Orf virus from the Genebank. For some reason the GFF3 file downloaded from Genebank has to be modified using some awk commands. You can find the script here. I want to acknowlege SMK for this solution.

After you convert your GFF3 file into new GFF3 format and save it as orf.gff3. Then it is pretty straightforward to create taxonomic database.

library (GenomicFeatures)
orf <- GenomicFeatures::makeTxDbFromGFF("content/post/orf.gff3",format="auto")
orf

Then you have to save this database,so that you can call it in future without recreating it again. We will use saveDb() to save database and loadDb() to load the database. We can use AnnotationDbi package fot this purpose. You will see a new file as orf in the working directory.

library (AnnotationDbi)
saveDb(orf,"orf")

As you can see we just created a taxonomic database for Orf virus. There are 130 CDS that have been extracted and you can use other functions as shown below.

First load the database and then try different funtions.

library (AnnotationDbi)
orf <- loadDb(file="orf")
#now check column names
columns(orf)

#extract genes
genes <- genes(orf,"GENEID")
head(genes)

#if you want to see the intergeninc regions
intergenic <- gaps (genes)

head (intergenic)

So, you can extract anything from this genome. Look at the columnames to see what kind of information you can extract.

This taxonomic database shows the postions where each gene or features are located within the genome. It does not actually contain any DNA sequence. To extract DNA sequence we need BSgenome package. Unfortunately, this virus or other viruses are not included in BSgenome package but the package allows you to make your own pakcage. This process is kown as “forging”. So, lets try to forge BSgenome for our Orf virus, so that we can work with DNA sequence of this virus.

This process involves a creating a src file that contains all the information about the genome. We will need fasta file of the genome, finally we will use command line tools to create a package. So, we will be creating a BSgenome package for Orf virus.

Steps :

First create a folder named Orf.

Inside that folder create a folder called src_seqdir.

Put your genome in fasta format inside that folder.

Create a plain text file and keep following information.Change the parameters as you like or that makes sense according to the genome that you are working with.


Package: OrfTxdb

Title: Full genome of Orf Virus Strain IA82

Description: Full genome of orf virus

Version: 1.0.0

Suggests: Orf.virus.NCBI

organism: Orfv

Author: Delhon et al.

common_name: Orf virus

provider: NCBI

provider_version: 1.0.0

release_date: Jan.2004

release_name: Orf virus strain OV-IA82

BSgenomeObjname: Orfv virus

source_url: https://www.ncbi.nlm.nih.gov/nuccore/AY386263.1

organism_biocview: AnnotationData, BSgenome

seqnames: c(“orfv”)

SrcDataFiles: orfv.fasta

seqfiles_suffix: .fasta

seqs_srcdir: C:_000


Then save this plain text file as seed_file.txt. Remember the package name. This will be the name of your package.

Convert this plain file into DCF format in R. First read the file and then save it as DCF format.

my_file <- read.dcf("content/post/seed_file.txt", fields = NULL, all = FALSE, keep.white = NULL)
write.dcf(my_file , file = "seed.dcf", append = FALSE, useBytes = FALSE,
          indent = 0.1 * getOption("width"),
          width = 0.9 * getOption("width"),
          keep.white = NULL)

We will get seed.dcf file as an output. Now we are ready to forge.

library (BSgenome)
#remove diectory if already exixts, since I will be running this code repeatedly
unlink(c("OrfTxdb"), recursive = TRUE, force = TRUE)

#now create BSgenome package
forgeBSgenomeDataPkg("seed.dcf")

It will create a new directory in the current directory. The name will be ORfTxdb as we supplied this name on the seed file.

Now we are ready to create package.


Go to the Terminal of the R Studio and navigate to the current folder. Enter following command

R CMD build OrfTxdb

Enter the name of the folder that was created above when by forgeBSgenomeDataPkg function. This will create a tarball file of our package.In my case a new tar file called OrfTxdb_1.0.0.tar was created.

Check if everything is okay with your package using following command in the terminal

R CMD check OrfTxdb_1.0.0.tar

If everything is correct, then install the package using follwing command:

R CMD install OrfTxdb_1.0.0.tar


Now you can go back to R studio console and call this package. For example I named my package as OrfTxdb’. Lets use some functions on this package :

library (OrfTxdb)
orf_genome <- OrfTxdb
orf_genome
## Orf virus genome:
## # organism: Orfv (Orf virus)
## # provider: NCBI
## # provider version: 1.0.0
## # release date: Jan.2004
## # release name: Orf virus strain OV-IA82
## # 1 sequences:
## #   orfv                                                                      
## # (use 'seqnames()' to see all the sequence names, use the '$' or '[[' operator
## # to access a given sequence)
organism(orf_genome)
## [1] "Orfv"
provider(orf_genome)
## [1] "NCBI"
#lets get the whome genome sequence
getSeq(orf_genome)
## 137241-letter DNAString object
## seq: CGAGAACGCGGACCAGGAGTTCCTGCGGGAGGAGCT...AGCTCCTCCCGCAGGAACTCCTGGTCCGCGTTCTCG
#lets extract the part of the geome
getSeq(orf_genome,start=24,end=58)
## 35-letter DNAString object
## seq: TGCGGGAGGAGCTACGGCAGAGGCTGGAACTGCTG

So, there are other cool things that can be done using genome features and BSgenome. For example, we can get postions of all the promoter sequences from genome features and use those values to extract promoter sequences from the BSgenome. I will write a post how to do that, but there are already tutorials available. For now, if we want to get the files and folders used in this tutorial you can get them here in my Github.

Lok Raj Joshi
Lok Raj Joshi
PhD Candidate

My research interests include vectored vaccine development, viral pathogenesis, diagnostic assay development,reverse genetics, viral bioinformatics.

Related