BioDT

Do all your bioinformatics from R, almost exclusively using data.tables and tools from that paradigm.

WORK IN PROGRESS

Represents common bioinformatics formats like fasta, bed, vcf, as data.tables with standardised columns, but which can be decorated or manipulated to the user's desire. Provides wrappers to seamlessly do common bioinformatics tasks that all read to and from data.tables. The user enjoys all the power and flexbility of data.table, and is spared the fiddliness and pain of jumping endlessly between primary processing in Linux and interactive exploration in R.

Includes:

• Wrappers for common linux bioinformatics tools for things like blast, lastz, bedtools, primer3, clustalo, cd-hit, tandemRepeatsFinder ... the list grows.
• Read/write functions (file ==> data.table; data.table ==> file) and converters (data.table ==> data.table).
• A very flexible and loose 'typing' system for data.tables, e.g. a 'seqDT' has columns "seqId" [character], "seq" ["character"], and as long as they have valid entries, all functions requiring a seqDT will work. A validation system checks data.tables for compliance with BioDT conventions and gives helpful feedback on how to fix any errors.
• SOON Interfaces e.g. to pluck out chunks of [bv]cf files as data.tables, using syntax like vcf.interface["chromosome",start,end]
• Various plotting and visualisation functions
• Various convenience functions for bioinformatics tasks that keep cropping up again and again (alleleFrequency(), He()[Effective Heterozygosity], setRefAlleleMostCommon(), ...), and just other very useful things (mostCommonThing(), )
• A very lightweight and powerful colour palette generator designed for layering custom plots in base::plot.
• SOON Documentation and vignettes.

BioDT is being developed by Tim Rabanus-Wallace from the CropGEM lab at the University of Melbourne.

Example:

# ${PATH} must include recent versions of blast, bedtools, awk, ...

# If needed. You will need to make sure `msa` and `Biostrings` are installed first using BiocManager. This will be fixed.
# remotes::install_github("https://github.com/mtrw/BioDT")

require(BioDT) # also loads data.table and magrittr

# Read in some 'seed' sequences to study from a file
seedSeqs <- readFasta("data/seed_seqs.fasta")

# find homologs in a reference genome
bl <- blast(subjectFname="data/genome.fasta",querySeqDT=seedSeqs)

# Remind yourself ... what constitutes a valid alignmentDT?
is_alignmentDT(show=TRUE)

# Plot results
bl[,{
  x <- pmean2(sStart,sEnd) # using BioDT::pmean2()
  y <- frank(sSeqId)
  col <- applyPalette( x=bitscore , colChain = palettePresets$mercedes ) #using BioDT's palette functions
  null_plot(x,y)
  points(x,y,col=col,pch=20,cex=2)
}]

# Filter for 5 best hits per query
setorder(bl,qSeqId,bitscore)
bestHits <- bl[,last(.SD,5),by=.(qSeqId)] # using data.table::last()

# Get coords of hits plus surrounding sequence
surround <- 500L
bestHitSurroundCoords <- alignmentDT2coordDT(bestHits,coordsOf = "subject")[,.(seqId,start=start-surround,end=end+surround)]

# Get the seqs of the hits plus surrounding context
bestHitSurroundSeq <- extractSeq(coordDT=bestHitSurroundCoords,fastaFname = "data/genome.fasta")

# Multiply align
msa <- MSA(bestHitSurroundSeq)

# Plot the alignment with custom colours
plotNucleotideAlignment( msa[,.(seqId,seq=alnSeq)] )

CITATION (to pre-release verion): Rabanus-Wallace, M.T. (2025) BioDT: All your bioinformatics done from R, using data.tables. doi.org/10.5281/zenodo.14927782