-
Notifications
You must be signed in to change notification settings - Fork 0
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
LDA
authored and
LDA
committed
Oct 22, 2024
1 parent
ef3f992
commit 8a6e0a7
Showing
3 changed files
with
199 additions
and
198 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -6,3 +6,4 @@ inst/doc | |
| /doc/ | ||
| /Meta/ | ||
| docs | ||
| .DS_Store | ||
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -1,181 +1,181 @@ | ||
| #' Normalization and assessment in one function | ||
| #' | ||
| #' @param object Enone object. | ||
| #' @param auto Whether to automatically select negative control, positive | ||
| #' evaluation and negative evaluation genes, default: TRUE. | ||
| #' @param return.norm Whether to return normalized counts in object. By default, | ||
| #' not return normalized counts for reducing memory costs. | ||
| #' @param n.neg.control Number of negative control genes for RUV normalization, default: 1000. | ||
| #' @param n.pos.eval Number of positive evaluation genes for wanted variation assessment, default: 500. | ||
| #' @param n.neg.eval Number of negative evaluation genes for unwanted variation assessment, default: 500. | ||
| #' @param neg.control Vector of negative control genes' id for RUV normalization, default: NULL. | ||
| #' @param pos.eval Vector of positive evaluation genes' id for wanted variation assessment, default: NULL. | ||
| #' @param neg.eval Vector of negative evaluation genes' id for unwanted variation assessment, default: NULL. | ||
| #' @param scaling.method Vector of scaling methods that are applied to the data. | ||
| #' Available methods are: \code{c("TC", "UQ", "TMM", "DESeq", "PossionSeq")}. | ||
| #' Select one or multiple methods. By default all scaling methods will be applied. | ||
| #' @param ruv.norm Whether to perform RUV normalization. | ||
| #' @param ruv.k The number of factors of unwanted variation to be estimated from the data, default: 1. | ||
| #' @param ruv.drop The number of singular values to drop in the estimation of | ||
| #' unwanted variation, default: 0. | ||
| #' @param eval.pam.k Integer or vector of integers indicates the number of | ||
| #' clusters for PAM clustering in performance evaluation, default: 2:6. | ||
| #' @param eval.pc.n Integer indicates the evaluation metrics will be calculated | ||
| #' in the first nth PCs, default: 3. | ||
| #' | ||
| #' @return Enone object. | ||
| #' @export | ||
| #' | ||
| #' @importFrom SummarizedExperiment rowData assay | ||
| #' @importFrom stats setNames | ||
| enONE <- function(object, | ||
| auto = TRUE, return.norm = FALSE, | ||
| n.neg.control = 1000, n.pos.eval = 500, n.neg.eval = 500, | ||
| neg.control = NULL, pos.eval = NULL, neg.eval = NULL, | ||
| scaling.method = c("TC", "UQ", "TMM", "DESeq", "PossionSeq"), | ||
| ruv.norm = TRUE, ruv.k = 1, ruv.drop = 0, | ||
| eval.pam.k = 2:6, eval.pc.n = 3) { | ||
|
|
||
| # check ruv.k, eval.pam.k and eval.pc.n are least than the number of samples | ||
| if (!all(c(ruv.k, eval.pam.k, eval.pc.n) < ncol(object))) { | ||
| stop("Number of `ruv.k`, `eval.pam.k` and `eval.pc.n` should not exceed the number of samples.") | ||
| } | ||
|
|
||
| # retrieve parameters from Enone object | ||
| bio.group <- object$condition | ||
| enrich.group <- object$enrich | ||
|
|
||
| if (!any(is.na(object$batch))) { | ||
| batch.group <- object$batch | ||
| } else { | ||
| batch.group <- NULL | ||
| } | ||
|
|
||
| spike.in.prefix <- object@parameter$spike.in.prefix | ||
| input.id <- object@parameter$input.id | ||
| enrich.id <- object@parameter$enrich.id | ||
| synthetic.id <- object@parameter$synthetic.id | ||
|
|
||
| # create group matrix | ||
| sc_mat <- CreateGroupMatrix(bio.group) | ||
| enrich_mat <- CreateGroupMatrix(enrich.group) | ||
|
|
||
| # get counts | ||
| data <- SummarizedExperiment::assay(object) | ||
| # sample counts | ||
| counts_nsp <- data[grep(paste(c(spike.in.prefix, synthetic.id), collapse = "|"), rownames(data), invert = TRUE),] | ||
| # spike-in counts | ||
| counts_sp <- data[grep(spike.in.prefix, rownames(data)),] | ||
|
|
||
| ## gene selection | ||
| if (auto) { | ||
| genes.ls <- GeneSelection(object, | ||
| n.neg.control = n.neg.control, | ||
| n.pos.eval = n.pos.eval, | ||
| n.neg.eval = n.neg.eval) | ||
|
|
||
| neg.control.set <- genes.ls[["NegControl"]] | ||
| pos.eval.set <- genes.ls[["PosEvaluation"]] | ||
| neg.eval.set <- genes.ls[["NegEvaluation"]] | ||
|
|
||
| } else { | ||
|
|
||
| if (!all(neg.control %in% rownames(data))) { | ||
| stop("`neg.control` are not presented in the rownames of count matrix.") | ||
| } else { | ||
| neg.control.set <- neg.control | ||
| } | ||
|
|
||
| if (!is.null(pos.eval) & !all(pos.eval %in% rownames(data))) { | ||
| stop("`pos.eval` are not presented in the rownames of count matrix.") | ||
| } else { | ||
| pos.eval.set <- pos.eval | ||
| } | ||
|
|
||
| if (!is.null(neg.eval) & !all(neg.eval %in% rownames(data))) { | ||
| stop("`neg.eval` are not presented in the rownames of count matrix.") | ||
| } else { | ||
| neg.eval.set <- neg.eval | ||
| } | ||
|
|
||
| } | ||
|
|
||
| # save gene set to object | ||
| SummarizedExperiment::rowData(object)$NegControl <- SummarizedExperiment::rowData(object)$GeneID %in% neg.control.set | ||
| SummarizedExperiment::rowData(object)$NegEvaluation <- SummarizedExperiment::rowData(object)$GeneID %in% neg.eval.set | ||
| SummarizedExperiment::rowData(object)$PosEvaluation <- SummarizedExperiment::rowData(object)$GeneID %in% pos.eval.set | ||
|
|
||
| ## apply normalization | ||
| cat("Apply normalization...\n") | ||
| norm.ls <- ApplyNormalization(data, | ||
| scaling.method = scaling.method, | ||
| ruv.norm = ruv.norm, ruv.k = ruv.k, ruv.drop = ruv.drop, | ||
| spike.in.prefix = spike.in.prefix, | ||
| synthetic.id = synthetic.id, | ||
| # below parameters are generated inside enONE function | ||
| control.idx = neg.control.set, | ||
| sc.idx = sc_mat, | ||
| enrich.idx = enrich_mat) | ||
|
|
||
| ## assessment | ||
| bio_group_index <- as.numeric(factor(bio.group, levels=unique(bio.group))) | ||
| enrich_group_index <- as.numeric(factor(enrich.group, levels=unique(enrich.group))) | ||
| if (!is.null(batch.group)) { | ||
| batch_group_index <- as.numeric(factor(batch.group, levels=unique(batch.group))) | ||
| } else { | ||
| batch_group_index <- NULL | ||
| } | ||
| cat("Perform assessment...\n") | ||
| norm.eval <- AssessNormalization(norm.ls, | ||
| eval.pam.k = eval.pam.k, | ||
| eval.pc.n = eval.pc.n, | ||
| batch.group = batch_group_index, | ||
| # below parameters are created inside function | ||
| bio.group = bio_group_index, | ||
| enrich.group = enrich_group_index, | ||
| pos.eval.set = pos.eval.set, | ||
| neg.eval.set = neg.eval.set) | ||
|
|
||
| ## save metrics to Enone object | ||
| object@enone_metrics <- norm.eval$metrics | ||
| ## save score to Enone object | ||
| object@enone_score <- norm.eval$score | ||
| ## add run parameter to Enone object | ||
| parameter.run <- list( | ||
| n.neg.control = n.neg.control, | ||
| n.pos.eval = n.pos.eval, | ||
| n.neg.eval = n.neg.eval, | ||
| scaling.method = scaling.method, | ||
| ruv.norm = ruv.norm, | ||
| ruv.k = ruv.k, | ||
| ruv.drop = ruv.drop, | ||
| eval.pam.k = eval.pam.k, | ||
| eval.pc.n = eval.pc.n, | ||
| auto = auto, | ||
| return.norm = return.norm | ||
| ) | ||
| object@parameter <- c(object@parameter, parameter.run) | ||
|
|
||
| # store normalization method names in object | ||
| norm.methods <- names(norm.ls) | ||
| object@counts$sample <- stats::setNames(vector("list", length(norm.methods)), nm = norm.methods) | ||
| object@enone_factor$sample <- stats::setNames(vector("list", length(norm.methods)), nm = norm.methods) | ||
|
|
||
| # store 'Raw' count matrix | ||
| Counts(object, slot = "sample", method = "Raw") <- counts_nsp | ||
| Counts(object, slot = "spike_in", method = "Raw") <- counts_sp | ||
|
|
||
| if (return.norm == TRUE) { | ||
| # store normalized counts | ||
| object@counts$sample <- lapply(norm.ls, function(i) { i$dataNorm }) | ||
| # store normalization factors | ||
| object@enone_factor$sample <- lapply(norm.ls, function(i) { i[c("normFactor", "adjustFactor")] }) | ||
| # rename factor slot | ||
| for (i in 1:length(object@enone_factor$sample)) { | ||
| names(object@enone_factor$sample[[i]]) <- c("normFactor", "adjustFactor") | ||
| } | ||
| } | ||
|
|
||
| validObject(object) | ||
| return(object) | ||
| } | ||
| #' Normalization and assessment in one function | ||
| #' | ||
| #' @param object Enone object. | ||
| #' @param auto Whether to automatically select negative control, positive | ||
| #' evaluation and negative evaluation genes, default: TRUE. | ||
| #' @param return.norm Whether to return normalized counts in object. By default, | ||
| #' not return normalized counts for reducing memory costs. | ||
| #' @param n.neg.control Number of negative control genes for RUV normalization, default: 1000. | ||
| #' @param n.pos.eval Number of positive evaluation genes for wanted variation assessment, default: 500. | ||
| #' @param n.neg.eval Number of negative evaluation genes for unwanted variation assessment, default: 500. | ||
| #' @param neg.control Vector of negative control genes' id for RUV normalization, default: NULL. | ||
| #' @param pos.eval Vector of positive evaluation genes' id for wanted variation assessment, default: NULL. | ||
| #' @param neg.eval Vector of negative evaluation genes' id for unwanted variation assessment, default: NULL. | ||
| #' @param scaling.method Vector of scaling methods that are applied to the data. | ||
| #' Available methods are: \code{c("TC", "UQ", "TMM", "DESeq", "PossionSeq")}. | ||
| #' Select one or multiple methods. By default all scaling methods will be applied. | ||
| #' @param ruv.norm Whether to perform RUV normalization. | ||
| #' @param ruv.k The number of factors of unwanted variation to be estimated from the data, default: 1. | ||
| #' @param ruv.drop The number of singular values to drop in the estimation of | ||
| #' unwanted variation, default: 0. | ||
| #' @param eval.pam.k Integer or vector of integers indicates the number of | ||
| #' clusters for PAM clustering in performance evaluation, default: 2:6. | ||
| #' @param eval.pc.n Integer indicates the evaluation metrics will be calculated | ||
| #' in the first nth PCs, default: 3. | ||
| #' | ||
| #' @return Enone object. | ||
| #' @export | ||
| #' | ||
| #' @importFrom SummarizedExperiment rowData assay | ||
| #' @importFrom stats setNames | ||
| enONE <- function(object, | ||
| auto = TRUE, return.norm = FALSE, | ||
| n.neg.control = 1000, n.pos.eval = 500, n.neg.eval = 500, | ||
| neg.control = NULL, pos.eval = NULL, neg.eval = NULL, | ||
| scaling.method = c("TC", "UQ", "TMM", "DESeq", "PossionSeq"), | ||
| ruv.norm = TRUE, ruv.k = 1, ruv.drop = 0, | ||
| eval.pam.k = 2:6, eval.pc.n = 3) { | ||
|
|
||
| # check ruv.k, eval.pam.k and eval.pc.n are least than the number of samples | ||
| if (!all(c(ruv.k, eval.pam.k, eval.pc.n) < ncol(object))) { | ||
| stop("Number of `ruv.k`, `eval.pam.k` and `eval.pc.n` should not exceed the number of samples.") | ||
| } | ||
|
|
||
| # retrieve parameters from Enone object | ||
| bio.group <- object$condition | ||
| enrich.group <- object$enrich | ||
|
|
||
| if (!any(is.na(object$batch))) { | ||
| batch.group <- object$batch | ||
| } else { | ||
| batch.group <- NULL | ||
| } | ||
|
|
||
| spike.in.prefix <- object@parameter$spike.in.prefix | ||
| input.id <- object@parameter$input.id | ||
| enrich.id <- object@parameter$enrich.id | ||
| synthetic.id <- object@parameter$synthetic.id | ||
|
|
||
| # create group matrix | ||
| sc_mat <- CreateGroupMatrix(bio.group) | ||
| enrich_mat <- CreateGroupMatrix(enrich.group) | ||
|
|
||
| # get counts | ||
| data <- SummarizedExperiment::assay(object) | ||
| # sample counts | ||
| counts_nsp <- data[grep(paste(c(spike.in.prefix, synthetic.id), collapse = "|"), rownames(data), invert = TRUE),] | ||
| # spike-in counts | ||
| counts_sp <- data[grep(spike.in.prefix, rownames(data)),] | ||
|
|
||
| ## gene selection | ||
| if (auto) { | ||
| genes.ls <- GeneSelection(object, | ||
| n.neg.control = n.neg.control, | ||
| n.pos.eval = n.pos.eval, | ||
| n.neg.eval = n.neg.eval) | ||
|
|
||
| neg.control.set <- genes.ls[["NegControl"]] | ||
| pos.eval.set <- genes.ls[["PosEvaluation"]] | ||
| neg.eval.set <- genes.ls[["NegEvaluation"]] | ||
|
|
||
| } else { | ||
|
|
||
| if (!all(neg.control %in% rownames(data))) { | ||
| stop("`neg.control` are not presented in the rownames of count matrix.") | ||
| } else { | ||
| neg.control.set <- neg.control | ||
| } | ||
|
|
||
| if (!is.null(pos.eval) & !all(pos.eval %in% rownames(data))) { | ||
| stop("`pos.eval` are not presented in the rownames of count matrix.") | ||
| } else { | ||
| pos.eval.set <- pos.eval | ||
| } | ||
|
|
||
| if (!is.null(neg.eval) & !all(neg.eval %in% rownames(data))) { | ||
| stop("`neg.eval` are not presented in the rownames of count matrix.") | ||
| } else { | ||
| neg.eval.set <- neg.eval | ||
| } | ||
|
|
||
| } | ||
|
|
||
| # save gene set to object | ||
| SummarizedExperiment::rowData(object)$NegControl <- SummarizedExperiment::rowData(object)$GeneID %in% neg.control.set | ||
| SummarizedExperiment::rowData(object)$NegEvaluation <- SummarizedExperiment::rowData(object)$GeneID %in% neg.eval.set | ||
| SummarizedExperiment::rowData(object)$PosEvaluation <- SummarizedExperiment::rowData(object)$GeneID %in% pos.eval.set | ||
|
|
||
| ## apply normalization | ||
| cat("Apply normalization...\n") | ||
| norm.ls <- ApplyNormalization(data, | ||
| scaling.method = scaling.method, | ||
| ruv.norm = ruv.norm, ruv.k = ruv.k, ruv.drop = ruv.drop, | ||
| spike.in.prefix = spike.in.prefix, | ||
| synthetic.id = synthetic.id, | ||
| # below parameters are generated inside enONE function | ||
| control.idx = neg.control.set, | ||
| sc.idx = sc_mat, | ||
| enrich.idx = enrich_mat) | ||
|
|
||
| ## assessment | ||
| bio_group_index <- as.numeric(factor(bio.group, levels=unique(bio.group))) | ||
| enrich_group_index <- as.numeric(factor(enrich.group, levels=unique(enrich.group))) | ||
| if (!is.null(batch.group)) { | ||
| batch_group_index <- as.numeric(factor(batch.group, levels=unique(batch.group))) | ||
| } else { | ||
| batch_group_index <- NULL | ||
| } | ||
| cat("Perform assessment...\n") | ||
| norm.eval <- AssessNormalization(norm.ls, | ||
| eval.pam.k = eval.pam.k, | ||
| eval.pc.n = eval.pc.n, | ||
| batch.group = batch_group_index, | ||
| # below parameters are created inside function | ||
| bio.group = bio_group_index, | ||
| enrich.group = enrich_group_index, | ||
| pos.eval.set = pos.eval.set, | ||
| neg.eval.set = neg.eval.set) | ||
|
|
||
| ## save metrics to Enone object | ||
| object@enone_metrics <- norm.eval$metrics | ||
| ## save score to Enone object | ||
| object@enone_score <- norm.eval$score | ||
| ## add run parameter to Enone object | ||
| parameter.run <- list( | ||
| n.neg.control = n.neg.control, | ||
| n.pos.eval = n.pos.eval, | ||
| n.neg.eval = n.neg.eval, | ||
| scaling.method = scaling.method, | ||
| ruv.norm = ruv.norm, | ||
| ruv.k = ruv.k, | ||
| ruv.drop = ruv.drop, | ||
| eval.pam.k = eval.pam.k, | ||
| eval.pc.n = eval.pc.n, | ||
| auto = auto, | ||
| return.norm = return.norm | ||
| ) | ||
| object@parameter <- c(object@parameter, parameter.run) | ||
|
|
||
| # store normalization method names in object | ||
| norm.methods <- names(norm.ls) | ||
| object@counts$sample <- stats::setNames(vector("list", length(norm.methods)), nm = norm.methods) | ||
| object@enone_factor$sample <- stats::setNames(vector("list", length(norm.methods)), nm = norm.methods) | ||
|
|
||
| # store 'Raw' count matrix | ||
| Counts(object, slot = "sample", method = "Raw") <- counts_nsp | ||
| Counts(object, slot = "spike_in", method = "Raw") <- counts_sp | ||
|
|
||
| if (return.norm == TRUE) { | ||
| # store normalized counts | ||
| object@counts$sample <- lapply(norm.ls, function(i) { i$dataNorm }) | ||
| # store normalization factors | ||
| object@enone_factor$sample <- lapply(norm.ls, function(i) { i[c("normFactor", "adjustFactor", "alpha")] }) | ||
| # rename factor slot | ||
| for (i in 1:length(object@enone_factor$sample)) { | ||
| names(object@enone_factor$sample[[i]]) <- c("normFactor", "adjustFactor", "alpha") | ||
| } | ||
| } | ||
|
|
||
| validObject(object) | ||
| return(object) | ||
| } |
Oops, something went wrong.