remap.data: map data to existing SOM

In SOMMD: Self Organising Maps for the Analysis of Molecular Dynamics Data

map data to existing SOM

Description

Assign new data to a pre-trained SOM

Usage

remap.data(SOM, X, add = FALSE)

Arguments

SOM	a trained SOM
X	a data set with the same number of features of the dataset used to train the SOM
add	whether to append the new data to the ones used to train the SOM

Value

An object of class "kohonen" with the new data mapped

Author(s)

Stefano Motta [e-mail]

Examples

#Read example SOM data
som_model <- readRDS(system.file("extdata", "SOM_HIFa.rds", package = "SOMMD"))
#Read a trajectory that was not used to train the som
trj_2 <- read.trj(trjfile = system.file("extdata", "HIF2a-MD-2.xtc", package = "SOMMD"),
  topfile = system.file("extdata", "HIF2a.gro", package = "SOMMD"))
#Read reference structure file
gro <- read.struct(system.file("extdata", "HIF2a.gro", package = "SOMMD"))
#Selection of the same intermolecular distances used to train the SOM
protein.sele <- which(gro$atom$resid!="020")
ligand.sele <- which(gro$atom$resid=="020")
heavy.atoms <- which(startsWith(gro$atom$elety, "H")==FALSE)
sele.dists <- native.cont(struct=gro, distance=0.6, mol.2=ligand.sele, atoms=heavy.atoms)
# Compute distances on new simulations (the same used for SOM training)
dist_2 <- calc.distances(trj_2, mol.2=ligand.sele, sele=sele.dists, atoms=heavy.atoms)
# Map new data on the existing SOM
som_model_2 <- remap.data(SOM=som_model, X=dist_2)

SOMMD documentation built on Oct. 2, 2024, 5:07 p.m.