The CaDrA package currently supports four scoring functions to search for subsets of genomic features that are likely associated with a specific outcome of interest (e.g., protein expression, pathway activity, etc.)
ks
)wilcox
)revealer
)knnmi
)correlation
)custom
)Below, we run candidate_search()
over the top 3 starting
features using each of the scoring functions described above.
Important Notes:
topn_eval()
is
equivalent to the new and recommended candidate_search()
functionbinary features matrix
also known as
Feature Set
(such as somatic mutations, copy number
alterations, chromosomal translocations, etc.) The 1/0 row vectors
indicate the presence/absence of ‘omics’ features in the samples. The
Feature Set
can be a matrix or an object of class
SummarizedExperiment from
SummarizedExperiment package)Input Scores
)
representing a functional response of interest (such as protein
expression, pathway activity, etc.)The simulated dataset, sim_FS
, comprises of 1000 genomic
features and 100 sample profiles. There are 10 left-skewed (i.e. True
Positive or TP) and 990 uniformly-distributed (i.e. True Null or TN)
features simulated in the dataset. Below is a heatmap of the first 100
features.
See ?ks_rowscore
for more details
ks_topn_l <- CaDrA::candidate_search(
FS = sim_FS,
input_score = sim_Scores,
method = "ks_pval", # Use Kolmogorov-Smirnov scoring function
method_alternative = "less", # Use one-sided hypothesis testing
weights = NULL, # If weights is provided, perform a weighted-KS test
search_method = "both", # Apply both forward and backward search
top_N = 3, # Evaluate top 3 starting points for the search
max_size = 10, # Allow at most 10 features in meta-feature matrix
do_plot = FALSE, # We will plot it AFTER finding the best hits
best_score_only = FALSE # Return all results from the search
)
# Now we can fetch the feature set of top N features that corresponded to the best scores over the top N search
ks_topn_best_meta <- topn_best(ks_topn_l)
# Visualize best meta-feature result
meta_plot(topn_best_list = ks_topn_best_meta)
See ?wilcox_rowscore
for more details
wilcox_topn_l <- CaDrA::candidate_search(
FS = sim_FS,
input_score = sim_Scores,
method = "wilcox_pval", # Use Wilcoxon Rank-Sum scoring function
method_alternative = "less", # Use one-sided hypothesis testing
search_method = "both", # Apply both forward and backward search
top_N = 3, # Evaluate top 3 starting points for the search
max_size = 10, # Allow at most 10 features in meta-feature matrix
do_plot = FALSE, # We will plot it AFTER finding the best hits
best_score_only = FALSE # Return all results from the search
)
# Now we can fetch the feature set of top N feature that corresponded to the best scores over the top N search
wilcox_topn_best_meta <- topn_best(topn_list = wilcox_topn_l)
# Visualize best meta-feature result
meta_plot(topn_best_list = wilcox_topn_best_meta)
See ?revealer_rowscore
for more details
revealer_topn_l <- CaDrA::candidate_search(
FS = sim_FS,
input_score = sim_Scores,
method = "revealer", # Use REVEALER's CMI scoring function
search_method = "both", # Apply both forward and backward search
top_N = 3, # Evaluate top 3 starting points for the search
max_size = 10, # Allow at most 10 features in meta-feature matrix
do_plot = FALSE, # We will plot it AFTER finding the best hits
best_score_only = FALSE # Return all results from the search
)
# Now we can fetch the ESet of top feature that corresponded to the best scores over the top N search
revealer_topn_best_meta <- topn_best(topn_list = revealer_topn_l)
# Visualize best meta-feature result
meta_plot(topn_best_list = revealer_topn_best_meta)
See ?knnmi_rowscore
for more details
knnmi_topn_l <- CaDrA::candidate_search(
FS = sim_FS,
input_score = sim_Scores,
method = "knnmi", # Use knnmi scoring function
search_method = "both", # Apply both forward and backward search
top_N = 3, # Evaluate top 3 starting points for the search
max_size = 10, # Allow at most 10 features in meta-feature matrix
do_plot = FALSE, # We will plot it AFTER finding the best hits
best_score_only = FALSE # Return all results from the search
)
# Now we can fetch the ESet of top feature that corresponded to the best scores over the top N search
knnmi_topn_best_meta <- topn_best(topn_list = knnmi_topn_l)
# Visualize best meta-feature result
meta_plot(topn_best_list = knnmi_topn_best_meta)
See ?corr_rowscore
for more details
corr_topn_l <- CaDrA::candidate_search(
FS = SummarizedExperiment::assay(sim_FS),
input_score = sim_Scores,
method = "correlation", # Use correlation scoring function
cmethod = "spearman", # Use spearman correlation scoring function
top_N = 3, # Evaluate top 3 starting points for the search
max_size = 10, # Allow at most 10 features in meta-feature matrix
do_plot = FALSE, # We will plot it AFTER finding the best hits
best_score_only = FALSE # Return all results from the search
)
# Now we can fetch the feature set of top N feature that corresponded to the best scores over the top N search
corr_topn_best_meta <- topn_best(topn_list = corr_topn_l)
# Visualize best meta-feature result
meta_plot(topn_best_list = corr_topn_best_meta)
See ?custom_rowscore
for more details
# A customized function using ks-test
customized_ks_rowscore <- function(FS, input_score, weights=NULL, meta_feature=NULL, alternative="less", metric="pval"){
metric <- match.arg(metric)
alternative <- match.arg(alternative)
# Check if meta_feature is provided
if(!is.null(meta_feature)){
# Getting the position of the known meta features
locs <- match(meta_feature, row.names(FS))
# Taking the union across the known meta features
if(length(locs) > 1) {
meta_vector <- as.numeric(ifelse(colSums(FS[locs,]) == 0, 0, 1))
}else{
meta_vector <- as.numeric(FS[locs, , drop=FALSE])
}
# Remove the meta features from the binary feature matrix
# and taking logical OR btw the remaining features with the meta vector
FS <- base::sweep(FS[-locs, , drop=FALSE], 2, meta_vector, `|`)*1
# Check if there are any features that are all 1s generated from
# taking the union between the matrix
# We cannot compute statistics for such features and thus they need
# to be filtered out
if(any(rowSums(FS) == ncol(FS))){
verbose("Features with all 1s generated from taking the matrix union ",
"will be removed before progressing...\n")
FS <- FS[rowSums(FS) != ncol(FS), , drop=FALSE]
# If no features remained after filtering, exist the function
if(nrow(FS) == 0) return(NULL)
}
}
# KS is a ranked-based method
# So we need to sort input_score from highest to lowest values
input_score <- sort(input_score, decreasing=TRUE)
# Re-order the matrix based on the order of input_score
FS <- FS[, names(input_score), drop=FALSE]
# Check if weights is provided
if(length(weights) > 0){
# Check if weights has any labels or names
if(is.null(names(weights)))
stop("The weights object must have names or labels that ",
"match the labels of input_score\n")
# Make sure its labels or names match the
# the labels of input_score
weights <- as.numeric(weights[names(input_score)])
}
# Get the alternative hypothesis testing method
alt_int <- switch(alternative, two.sided=0L, less=1L, greater=-1L, 1L)
# Compute the ks statistic and p-value per row in the matrix
ks <- .Call(ks_genescore_mat_, FS, weights, alt_int)
# Obtain score statistics from KS method
# Change values of 0 to the machine lowest value to avoid taking -log(0)
stat <- ks[1,]
# Obtain p-values from KS method
# Change values of 0 to the machine lowest value to avoid taking -log(0)
pval <- ks[2,]
pval[which(pval == 0)] <- .Machine$double.xmin
# Compute the scores according to the provided metric
scores <- ifelse(rep(metric, nrow(FS)) %in% "pval", -log(pval), stat)
names(scores) <- rownames(FS)
return(scores)
}
# Search for best features using a custom-defined function
custom_topn_l <- CaDrA::candidate_search(
FS = SummarizedExperiment::assay(sim_FS),
input_score = sim_Scores,
method = "custom", # Use custom scoring function
custom_function = customized_ks_rowscore, # Use a customized scoring function
custom_parameters = NULL, # Additional parameters to pass to custom_function
weights = NULL, # If weights is provided, perform a weighted test
search_method = "both", # Apply both forward and backward search
top_N = 3, # Evaluate top 3 starting points for the search
max_size = 10, # Allow at most 10 features in meta-feature matrix
do_plot = FALSE, # We will plot it AFTER finding the best hits
best_score_only = FALSE # Return all results from the search
)
# Now we can fetch the feature set of top N feature that corresponded to the best scores over the top N search
custom_topn_best_meta <- topn_best(topn_list = custom_topn_l)
# Visualize best meta-feature result
CaDrA::meta_plot(topn_best_list = custom_topn_best_meta)
For validation purposes, compare the custom and built-in function.
topn_res <- CaDrA::candidate_search(
FS = sim_FS,
input_score = sim_Scores,
method = "ks_pval", # Use Kolmogorov-Smirnov scoring function
method_alternative = "less", # Use one-sided hypothesis testing
weights = NULL, # If weights is provided, perform a weighted-KS test
search_method = "both", # Apply both forward and backward search
top_N = 3, # Evaluate top 7 starting points for each search
max_size = 10, # Maximum size a meta-feature matrix can extend to
do_plot = FALSE, # Plot after finding the best features
best_score_only = FALSE # Return all results from the search
)
## Fetch the meta-feature set corresponding to its best scores over top N features searches
topn_best_meta <- topn_best(topn_res)
# Visualize the best results with the meta-feature plot
meta_plot(topn_best_list = topn_best_meta)
sessionInfo()
R version 4.4.2 (2024-10-31)
Platform: x86_64-pc-linux-gnu
Running under: Ubuntu 24.04.1 LTS
Matrix products: default
BLAS: /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3
LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/libopenblasp-r0.3.26.so; LAPACK version 3.12.0
locale:
[1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C
[3] LC_TIME=en_US.UTF-8 LC_COLLATE=C
[5] LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8
[7] LC_PAPER=en_US.UTF-8 LC_NAME=C
[9] LC_ADDRESS=C LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
time zone: Etc/UTC
tzcode source: system (glibc)
attached base packages:
[1] stats4 stats graphics grDevices utils datasets methods
[8] base
other attached packages:
[1] CaDrA_1.5.0 pheatmap_1.0.12
[3] SummarizedExperiment_1.37.0 Biobase_2.67.0
[5] GenomicRanges_1.59.1 GenomeInfoDb_1.43.2
[7] IRanges_2.41.1 S4Vectors_0.45.2
[9] BiocGenerics_0.53.3 generics_0.1.3
[11] MatrixGenerics_1.19.0 matrixStats_1.4.1
[13] testthat_3.2.1.1 devtools_2.4.5
[15] usethis_3.1.0 rmarkdown_2.29
loaded via a namespace (and not attached):
[1] bitops_1.0-9 tcltk_4.4.2 remotes_2.5.0
[4] rlang_1.1.4 magrittr_2.0.3 compiler_4.4.2
[7] vctrs_0.6.5 reshape2_1.4.4 stringr_1.5.1
[10] profvis_0.4.0 pkgconfig_2.0.3 crayon_1.5.3
[13] fastmap_1.2.0 XVector_0.47.0 ellipsis_0.3.2
[16] labeling_0.4.3 caTools_1.18.3 utf8_1.2.4
[19] promises_1.3.2 sessioninfo_1.2.2 UCSC.utils_1.3.0
[22] purrr_1.0.2 xfun_0.49 zlibbioc_1.52.0
[25] cachem_1.1.0 jsonlite_1.8.9 later_1.4.1
[28] DelayedArray_0.33.2 parallel_4.4.2 R6_2.5.1
[31] RColorBrewer_1.1-3 bslib_0.8.0 stringi_1.8.4
[34] pkgload_1.4.0 brio_1.1.5 jquerylib_0.1.4
[37] Rcpp_1.0.13-1 iterators_1.0.14 knitr_1.49
[40] R.utils_2.12.3 httpuv_1.6.15 Matrix_1.7-1
[43] R.cache_0.16.0 rstudioapi_0.17.1 abind_1.4-8
[46] yaml_2.3.10 doParallel_1.0.17 gplots_3.2.0
[49] codetools_0.2-20 miniUI_0.1.1.1 misc3d_0.9-1
[52] pkgbuild_1.4.5 lattice_0.22-6 tibble_3.2.1
[55] plyr_1.8.9 shiny_1.9.1 withr_3.0.2
[58] evaluate_1.0.1 desc_1.4.3 urlchecker_1.0.1
[61] pillar_1.9.0 KernSmooth_2.23-24 foreach_1.5.2
[64] rprojroot_2.0.4 ggplot2_3.5.1 munsell_0.5.1
[67] scales_1.3.0 gtools_3.9.5 xtable_1.8-4
[70] glue_1.8.0 ppcor_1.1 maketools_1.3.1
[73] tools_4.4.2 sys_3.4.3 buildtools_1.0.0
[76] fs_1.6.5 grid_4.4.2 knnmi_1.0
[79] colorspace_2.1-1 GenomeInfoDbData_1.2.13 cli_3.6.3
[82] fansi_1.0.6 S4Arrays_1.7.1 gtable_0.3.6
[85] R.methodsS3_1.8.2 sass_0.4.9 digest_0.6.37
[88] SparseArray_1.7.2 farver_2.1.2 htmlwidgets_1.6.4
[91] memoise_2.0.1 htmltools_0.5.8.1 R.oo_1.27.0
[94] lifecycle_1.0.4 httr_1.4.7 mime_0.12
[97] MASS_7.3-61