johrpan/geposan
1
0
Fork 0
mirror of https://github.com/johrpan/geposan.git synced 2025-10-26 10:47:25 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions

neural: Always use position data

Browse source
This commit is contained in:
Elias Projahn 2021-11-14 19:27:11 +01:00
parent 4b1ab9fc9a
commit 1af138bf6f
2 changed files with 14 additions and 27 deletions
Download patch file Download diff file Expand all files Collapse all files

5
R/analyze.R
View file

@ -41,10 +41,7 @@ analyze <- function(preset, progress = NULL) {
correlation(..., use_positions = TRUE) correlation(..., use_positions = TRUE)
}, },
"proximity" = proximity, "proximity" = proximity,
"neural" = neural, "neural" = neural
"neural_positions" = function(...) {
neural(..., use_positions = TRUE)
}
) )
results <- cached("analysis", preset, { results <- cached("analysis", preset, {

36
R/neural.R
View file

@ -1,17 +1,14 @@
# Find genes by training a neural network on reference position data. # Find genes by training a neural network on reference position data.
# #
# @param seed A seed to get reproducible results. # @param seed A seed to get reproducible results.
neural <- function(preset, neural <- function(preset, progress = NULL, seed = 448077) {
use_positions = FALSE,
progress = NULL,
seed = 448077) {
species_ids <- preset$species_ids species_ids <- preset$species_ids
gene_ids <- preset$gene_ids gene_ids <- preset$gene_ids
reference_gene_ids <- preset$reference_gene_ids reference_gene_ids <- preset$reference_gene_ids
cached( cached(
"neural", "neural",
c(species_ids, gene_ids, reference_gene_ids, use_positions), c(species_ids, gene_ids, reference_gene_ids),
{ # nolint { # nolint
set.seed(seed) set.seed(seed)
gene_count <- length(gene_ids) gene_count <- length(gene_ids)
@ -29,27 +26,20 @@ neural <- function(preset,
# enough data. # enough data.
species_ids_included <- NULL species_ids_included <- NULL
# Make a column containing distance data for each species. # Make a column containing positions for each species.
for (species_id in species_ids) { for (species_id in species_ids) {
species_data <- if (use_positions) { species_data <- distances[
setnames(distances[ species == species_id,
species == species_id, .(gene, position)
.(gene, position) ]
], "position", "distance")
} else {
distances[
species == species_id,
.(gene, distance)
]
}
# Only include species with at least 25% known values. # Only include species with at least 25% known values.
species_distances <- stats::na.omit(species_data) species_data <- stats::na.omit(species_data)
if (nrow(species_distances) >= 0.25 * gene_count) { if (nrow(species_data) >= 0.25 * gene_count) {
species_ids_included <- c(species_ids_included, species_id) species_ids_included <- c(species_ids_included, species_id)
data <- merge(data, species_distances, all.x = TRUE) data <- merge(data, species_data, all.x = TRUE)
# Replace missing data with mean values. The neural network # Replace missing data with mean values. The neural network
# can't handle NAs in a meaningful way. Choosing extreme # can't handle NAs in a meaningful way. Choosing extreme
@ -58,11 +48,11 @@ neural <- function(preset,
# However, this will of course lessen the significance of # However, this will of course lessen the significance of
# the results. # the results.
mean_distance <- round(species_distances[, mean(distance)]) mean_position <- round(species_data[, mean(position)])
data[is.na(distance), distance := mean_distance] data[is.na(position), position := mean_position]
# Name the new column after the species. # Name the new column after the species.
setnames(data, "distance", species_id) setnames(data, "position", species_id)
} }
} }