Profile-based protein representation derived by PSSM (Position-Specific Scoring Matrix) and auto cross covariance

This function calculates the feature vector based on the PSSM by running PSI-Blast and auto cross covariance tranformation.

Usage

extractPSSMAcc(pssmmat, lag)

Arguments

pssmmat: The PSSM computed by extractPSSM.
lag: The lag parameter. Must be less than the number of amino acids in the sequence (i.e. the number of columns in the PSSM matrix).

Value

A length lag * 20^2 named numeric vector, the element names are derived by the amino acid name abbreviation (crossed amino acid name abbreviation) and lag index.

References

Wold, S., Jonsson, J., Sjorstrom, M., Sandberg, M., & Rannar, S. (1993). DNA and peptide sequences and chemical processes multivariately modelled by principal component analysis and partial least-squares projections to latent structures. Analytica chimica acta, 277(2), 239–253.

Author

Nan Xiao <https://nanx.me>

Examples

if (Sys.which("makeblastdb") == "" | Sys.which("psiblast") == "") {
  cat("Cannot find makeblastdb or psiblast. Please install NCBI Blast+")
} else {
  x <- readFASTA(system.file(
    "protseq/P00750.fasta",
    package = "protr"
  ))[[1]]
  dbpath <- tempfile("tempdb", fileext = ".fasta")
  invisible(file.copy(from = system.file(
    "protseq/Plasminogen.fasta",
    package = "protr"
  ), to = dbpath))

  pssmmat <- extractPSSM(seq = x, database.path = dbpath)
  pssmacc <- extractPSSMAcc(pssmmat, lag = 3)
  tail(pssmacc)
}
#> Cannot find makeblastdb or psiblast. Please install NCBI Blast+