Compute the KL divergence between two given PSSMs

The KL divergence is computed by shifting the shorter PSSM along the longer one and computing the divergence at each position. The minimum divergence is returned.

Usage

pssm_diff(pssm1, pssm2, prior = 0.01)

Arguments

pssm1: first PSSM matrix or data frame
pssm2: second PSSM matrix or data frame
prior: a prior probability for each nucleotide.

Value

KL divergence between the two PSSMs (lower values indicate more similar PSSMs)

Examples

if (FALSE) { # \dontrun{
res1 <- regress_pwm(cluster_sequences_example, cluster_mat_example[, 1])
pssm_diff(res1$pssm, JASPAR_motifs[JASPAR_motifs$motif == "HNF1A", ])
} # }

# Compare KL divergence between two motifs from MOTIF_DB
pssm_diff(as.matrix(MOTIF_DB["HOMER.GATA3_2"]), as.matrix(MOTIF_DB["JASPAR.CDX1"]))
#> [1] 8.327341

# Lower values indicate more similar motifs
pssm_diff(as.matrix(MOTIF_DB["HOMER.GATA3_2"]), as.matrix(MOTIF_DB["JASPAR.GATA3"]))
#> [1] 9.37645