diff --git a/Example data files/Example_3_BIOPLEX_MFI-WITH-BEAD-COUNTS_XLSX.xlsx b/Example data files/Example_3_BIOPLEX_MFI-WITH-BEAD-COUNTS_XLSX.xlsx
new file mode 100644
index 0000000000000000000000000000000000000000..6ef56a6d0c5f4e685b57c2bd8fdedb99cbf12ad6
Binary files /dev/null and b/Example data files/Example_3_BIOPLEX_MFI-WITH-BEAD-COUNTS_XLSX.xlsx differ
diff --git a/Example data files/Example_3_plate_layout_XLSX.xlsx b/Example data files/Example_3_plate_layout_XLSX.xlsx
new file mode 100644
index 0000000000000000000000000000000000000000..ee9bfeac6a710506c8f7b05ebc31005fdc5c3556
Binary files /dev/null and b/Example data files/Example_3_plate_layout_XLSX.xlsx differ
diff --git a/Example data files/Example_4_BIOPLEX_MFI-WITHOUT-BEAD-COUNTS_XLSX.xlsx b/Example data files/Example_4_BIOPLEX_MFI-WITHOUT-BEAD-COUNTS_XLSX.xlsx
new file mode 100644
index 0000000000000000000000000000000000000000..b21b3433f3f2fa83ab48f91a673f7d2f41596951
Binary files /dev/null and b/Example data files/Example_4_BIOPLEX_MFI-WITHOUT-BEAD-COUNTS_XLSX.xlsx differ
diff --git a/Example data files/Example_4_plate_layout_XLSX.xlsx b/Example data files/Example_4_plate_layout_XLSX.xlsx
new file mode 100644
index 0000000000000000000000000000000000000000..02b3c156612d7905d343205b67bda64ee790e038
Binary files /dev/null and b/Example data files/Example_4_plate_layout_XLSX.xlsx differ
diff --git a/FUNCTIONS.R b/FUNCTIONS.R
index 31371cb023d2477cdf032a6e7962a7f2e5eff203..e3bdc23e7562a437fcf43dc14556a103a356eb7a 100644
--- a/FUNCTIONS.R
+++ b/FUNCTIONS.R
@@ -41,7 +41,7 @@ runRelativeAntibodyUnits = function(fname1, fname2, MFI_CSV, MFI_N_ANTIGENS, TEM
## converts the median fluorescence intensity (MFI) values into relative antibody units. ##
## ##
## Input: ##
- ## - Luminex-200 or Magpix output file (required) ##
+ ## - Luminex-200 or Magpix output file or BioPlex raw MFI output file(required) ##
## - Plate layout with bleedcode information (optional) ##
## ##
## Output: ##
@@ -56,7 +56,9 @@ runRelativeAntibodyUnits = function(fname1, fname2, MFI_CSV, MFI_N_ANTIGENS, TEM
#### READ DATA
##########################################################################################################
- # ## The first 41 rows are not relevant and are not imported into R.
+ ## The first rows are headers not relevant and are quickly discarded into R.
+ ## The number of rows to discard depends on platform used to generate the input.
+
if(MFI_CSV){
@@ -101,7 +103,7 @@ runRelativeAntibodyUnits = function(fname1, fname2, MFI_CSV, MFI_N_ANTIGENS, TEM
dim(L)
# MFI values as numeric
- L[,-which(colnames(L) %in% c("Location","Sample","Total Events","TotalEvents"))] = lapply(L[,-which(colnames(L) %in% c("Location","Sample","Total Events"))], as.numeric)
+ L[,-which(colnames(L) %in% c("Location","Sample","Total Events","TotalEvents"))] = lapply(L[,-which(colnames(L) %in% c("Location","Sample","Total Events","TotalEvents"))], as.numeric)
## Load the counts to check for run quality control
C = L_full[(count_row_number+1):(count_row_number+1+nrow(L)),1:(3+as.integer(MFI_N_ANTIGENS))]
@@ -133,70 +135,138 @@ runRelativeAntibodyUnits = function(fname1, fname2, MFI_CSV, MFI_N_ANTIGENS, TEM
# Read
L_full <- as.data.frame(read_excel(fname1))
- # Identify rows
- median_row_number <- which(L_full$xPONENT == "Median")
- count_row_number <- which(L_full$xPONENT == "Count")
- endcount_row_number <- which(L_full$xPONENT == "Avg Net MFI")
-
- # Load Excel file
- L <- as.data.frame(read_excel(fname1, skip = median_row_number+1, col_types = "text"))
+ # Guess the platform from the header structure
+ # This should be improved in the future (made into a radio button instead ?)
+ PLATFORM <- ifelse(any(grepl("MAGPIX", colnames(L_full))), "Magpix",
+ ifelse(any(grepl("Reader Serial Number", L_full[, 1])), "Bioplex", "Unknown"))
- ## Find all blank rows (i.e. rows that are all NA).
- ## Then keep rows preceding the first blank row.
- blank.row.number <- which(rowSums(is.na(L)) == length(names(L)))[1]
- if(is.na(blank.row.number)){
- L = L
- }else{
- L <- L[1:(blank.row.number-1),]
+ ## Magpix data input detected
+ if (PLATFORM == "Magpix") {
+ cat("Guessed MFI input file from a Magpix platform")
+
+ # Identify rows
+ median_row_number <- which(L_full$xPONENT == "Median")
+ count_row_number <- which(L_full$xPONENT == "Count")
+ endcount_row_number <- which(L_full$xPONENT == "Avg Net MFI")
+
+ # Load Excel file
+ L <- as.data.frame(read_excel(fname1, skip = median_row_number+1, col_types = "text"))
+
+ ## Find all blank rows (i.e. rows that are all NA).
+ ## Then keep rows preceding the first blank row.
+ blank.row.number <- which(rowSums(is.na(L)) == length(names(L)))[1]
+ if(is.na(blank.row.number)){
+ L = L
+ }else{
+ L <- L[1:(blank.row.number-1),]
+ }
+
+ ## Exclude column that corresponds to "Total events"
+ L <- L[, !(colnames(L) %in% c("Total Events","TotalEvents"))]
+
+ # Antigen names clean-up
+ colnames(L) = gsub("\\..*", "", colnames(L))
+ # Remove any values in (parentheses)
+ colnames(L) = gsub("\\s*\\([^\\)]+\\)","", colnames(L))
+ # Remove spaces
+ colnames(L) = gsub(" ","", colnames(L))
+
+ dim(L)
+
+ # Change "NaN" to 0s
+ L <- L %>% mutate_all(funs(gsub("NaN", 0, .)))
+
+ # MFI values as numeric
+ L[,-which(colnames(L) %in% c("Location","Sample","Total Events","TotalEvents"))] = lapply(L[,-which(colnames(L) %in% c("Location","Sample","Total Events","TotalEvents"))], as.numeric)
+
+ C <- as.data.frame(read_excel(fname1, skip = count_row_number+1, col_types = "text"))
+
+ ## Find all blank rows (i.e. rows that are all NA).
+ ## Then keep rows preceding the first blank row.
+ blank.row.number <- which(rowSums(is.na(C)) == length(names(C)))[1]
+ if(is.na(blank.row.number)){
+ C = C
+ }else{
+ C <- C[1:(blank.row.number-1),]
+ }
+
+ ## Exclude column that corresponds to "Total events"
+ C <- C[, !(colnames(C) %in% c("Total Events","TotalEvents"))]
+
+ # Antigen names clean-up
+ colnames(C) = gsub("\\..*", "", colnames(C))
+ # Remove any values in (parentheses)
+ colnames(C) = gsub("\\s*\\([^\\)]+\\)","", colnames(C))
+ # Remove spaces
+ colnames(C) = gsub(" ","", colnames(C))
+ dim(C)
+
+ ## Save the MFI values for the blank sample(s) for run quality control
+ B <- L %>% filter(grepl(c("^B"), Sample, ignore.case = T))
+ dim(B)
+
+ ## Save the MFI values for the standards for run quality control
+ S <- L %>% filter(grepl("^S", Sample, ignore.case = T))
+ dim(S) #there should be 10 rows
}
- ## Exclude column that corresponds to "Total events"
- L <- L[, !(colnames(L) %in% c("Total Events","TotalEvents"))]
-
- # Antigen names clean-up
- colnames(L) = gsub("\\..*", "", colnames(L))
- # Remove any values in (parentheses)
- colnames(L) = gsub("\\s*\\([^\\)]+\\)","", colnames(L))
- # Remove spaces
- colnames(L) = gsub(" ","", colnames(L))
-
- dim(L)
-
- # Change "NaN" to 0s
- L <- L %>% mutate_all(funs(gsub("NaN", 0, .)))
-
- # MFI values as numeric
- L[,-which(colnames(L) %in% c("Location","Sample","Total Events","TotalEvents"))] = lapply(L[,-which(colnames(L) %in% c("Location","Sample","Total Events"))], as.numeric)
-
- C <- as.data.frame(read_excel(fname1, skip = count_row_number+1, col_types = "text"))
-
- ## Find all blank rows (i.e. rows that are all NA).
- ## Then keep rows preceding the first blank row.
- blank.row.number <- which(rowSums(is.na(C)) == length(names(C)))[1]
- if(is.na(blank.row.number)){
- C = C
- }else{
- C <- C[1:(blank.row.number-1),]
+ ## BioPlex data input detected
+ else if (PLATFORM == "Bioplex") {
+ cat("Guessed MFI input file from a BioPlex platform")
+
+ # In BioPlex outputs, MFI and bead counts are in the same cell with format MFI (counts).
+ # Note that the bead counts may not be present, in that case only the MFI is reported.
+
+ # Identify rows (-1 because we match the header, contrary as one row before in Magpix)
+ median_row_number <- which(L_full[, 1] == "Well") - 1
+
+ # Load again Excel file reading only after the 1st row of interest
+ L <- as.data.frame(read_excel(fname1, skip = median_row_number+1, col_types = "text")) %>%
+ dplyr::rename(Location = Well,
+ Sample = Type) %>%
+ # Pad the Location column to match the syntax used by Luminex outputs,
+ # so that future regex will work regardless
+ dplyr::mutate(Location = paste0(1:n(), "(1,", Location, ")"))
+
+ # Remove columns that are not MFI readings or sample code / location
+ L <- L[, !(colnames(L) %in% c("Region", "Gate", "Total", "% Agg Beads", "Sampling Errors"))]
+
+ # Antigen names clean-up
+ colnames(L) = gsub("\\..*", "", colnames(L))
+ # Remove any values in (parentheses)
+ colnames(L) = gsub("\\s*\\([^\\)]+\\)","", colnames(L))
+ # Remove spaces
+ colnames(L) = gsub(" ","", colnames(L))
+
+ dim(L)
+
+ # Change "NaN" to 0s
+ L <- L %>% mutate_all(funs(gsub("NaN", 0, .)))
+
+ # Bead counts, when available, will be extracted from the table
+ C <- L %>%
+ dplyr::mutate(dplyr::across(!c(Location, Sample), gsub, pattern = "^(\\d+(?:\\.\\d+)?)(\\s*\\((\\d+(?:\\.\\d+)?)\\))?$", replacement = "\\3")) %>%
+ # Set to NA in case bead counts were not available
+ dplyr::mutate(dplyr::across(!c(Location, Sample), gsub, pattern = "^$", replacement = NA))
+
+ dim(C)
+
+ # Remove bead counts from L when they have been isolated in C
+ L <- L %>%
+ dplyr::mutate(dplyr::across(!c(Location, Sample), gsub, pattern = "^(\\d+(?:\\.\\d+)?)(\\s*\\((\\d+(?:\\.\\d+)?)\\))?$", replacement = "\\1"))
+
+ ## Save the MFI values for the blank sample(s) for run quality control
+ B <- L %>% filter(grepl(c("^B"), Sample, ignore.case = T))
+ dim(B)
+
+ ## Save the MFI values for the standards for run quality control
+ S <- L %>% filter(grepl("^S", Sample, ignore.case = T))
+ dim(S) #there should be 10 rows
}
- ## Exclude column that corresponds to "Total events"
- C <- C[, !(colnames(C) %in% c("Total Events","TotalEvents"))]
-
- # Antigen names clean-up
- colnames(C) = gsub("\\..*", "", colnames(C))
- # Remove any values in (parentheses)
- colnames(C) = gsub("\\s*\\([^\\)]+\\)","", colnames(C))
- # Remove spaces
- colnames(C) = gsub(" ","", colnames(C))
- dim(C)
-
- ## Save the MFI values for the blank sample(s) for run quality control
- B <- L %>% filter(grepl(c("^B"), Sample, ignore.case = T))
- dim(B)
-
- ## Save the MFI values for the standards for run quality control
- S <- L %>% filter(grepl("^S", Sample, ignore.case = T))
- dim(S) #there should be 10 rows
+ else {
+ stop("Execution halted: the platform used to generate the input could not be determined.")
+ }
}
@@ -214,9 +284,9 @@ runRelativeAntibodyUnits = function(fname1, fname2, MFI_CSV, MFI_N_ANTIGENS, TEM
dplyr::filter(grepl("^\\d+$",count)) %>%
dplyr::mutate(count = as.numeric(count),
warning = case_when(
- count<15~1,
- count>=15~0
- )) %>%
+ count<15~1,
+ count>=15~0
+ )) %>%
dplyr::select(Location, warning) %>%
dplyr::group_by(Location) %>%
dplyr::summarise(sum = sum(warning)) %>%
@@ -246,8 +316,8 @@ runRelativeAntibodyUnits = function(fname1, fname2, MFI_CSV, MFI_N_ANTIGENS, TEM
## Protein name list obtained after removing variable names: "Location" and "Sample"
proteins <- names(L[, -c(1:2)]); proteins
- ## Add new variable to data frame to indicate the first letter of sample type ("B", "C", "S", "U","N")
- ## "B" = blank, "C"=control, "S"=standard (dilution of the pool), "U"=sample, "N"=negative control
+ ## Add new variable to data frame to indicate the first letter of sample type ("B", "C", "S", "U" or "X","N")
+ ## "B" = blank, "C"=control, "S"=standard (dilution of the pool), "U" or "X"=sample, "N"=negative control
L$type.letter <- substr(L$Sample, start=1, stop=1)
dilution <- c(1/50, 1/100, 1/200, 1/400, 1/800, 1/1600, 1/3200, 1/6400, 1/12800, 1/25600)
dilution.scaled <- dilution*25600; dilution.scaled
@@ -303,7 +373,7 @@ runRelativeAntibodyUnits = function(fname1, fname2, MFI_CSV, MFI_N_ANTIGENS, TEM
for (r in 1:nrow(L)){
results <- NULL
- if (L$type.letter[r]=="U"){
+ if (L$type.letter[r] %in% c("U", "X")){
mfi.X <- as.numeric(L[r, i])
y <- log(mfi.X)
@@ -367,7 +437,7 @@ runRelativeAntibodyUnits = function(fname1, fname2, MFI_CSV, MFI_N_ANTIGENS, TEM
# Make all columns after 1st 4 numeric
results.df.wide[,5:ncol(results.df.wide)] = lapply(results.df.wide[,5:ncol(results.df.wide)], as.character)
results.df.wide[,5:ncol(results.df.wide)] = lapply(results.df.wide[,5:ncol(results.df.wide)], as.numeric)
-
+
##########################################################################################################
#### OUTPUT
@@ -428,7 +498,7 @@ getRelativeAntibodyUnits = function(RAW_MFI_FILE_NAME, RAW_MFI_FILE_PATH,
cat("******************Running RAU function******************\n")
- # Get Luminex-MagPix full file path
+ # Get Luminex-MagPix-BioPlex full file path
fname1 = RAW_MFI_FILE_PATH
# Extract working directly of this file
@@ -461,46 +531,53 @@ getRelativeAntibodyUnits = function(RAW_MFI_FILE_NAME, RAW_MFI_FILE_PATH,
bead_counts = results[[4]]
MFI_RAU_results = results[[5]]
model_results = results[[6]]
+ plot_stdcurve <- plot_counts <- plot_blank <- plots_model <- all_model_plots <- NULL
## Plot standard curve raw MFI
- plot_stdcurve <- std_curve %>%
- dplyr::select(-Location) %>%
- dplyr::mutate(Sample = c("S1", "S2", "S3", "S4", "S5", "S6", "S7", "S8", "S9", "S10"),
- Sample_dilution = factor(paste0(Sample,": ",dilution_plot),levels=paste0(Sample,": ",dilution_plot))) %>%
- tidyr::pivot_longer(-c(Sample,dilution,dilution_plot,Sample_dilution), names_to = "protein", values_to = "MFI") %>%
- # mutate(Sample = factor(Sample, c("S1", "S2", "S3", "S4", "S5", "S6", "S7", "S8", "S9", "S10"))) %>%
- dplyr::mutate(MFI = as.numeric(MFI)) %>%
- ggplot(aes(x = Sample_dilution, y = MFI, color = protein, group = protein)) +
- geom_point() +
- geom_line() +
- scale_y_log10(breaks = c(0, 10, 100, 1000, 10000)) +
- labs(x = "standard curve",
- y = "log(MFI)") +
- facet_wrap(~protein) +
- theme_bw() +
- theme(axis.text.x = element_text(angle = 90, hjust = 1))
+ if (nrow(std_curve) > 0) {
+ plot_stdcurve <- std_curve %>%
+ dplyr::select(-Location) %>%
+ dplyr::mutate(Sample = c("S1", "S2", "S3", "S4", "S5", "S6", "S7", "S8", "S9", "S10"),
+ Sample_dilution = factor(paste0(Sample,": ",dilution_plot),levels=paste0(Sample,": ",dilution_plot))) %>%
+ tidyr::pivot_longer(-c(Sample,dilution,dilution_plot,Sample_dilution), names_to = "protein", values_to = "MFI") %>%
+ # mutate(Sample = factor(Sample, c("S1", "S2", "S3", "S4", "S5", "S6", "S7", "S8", "S9", "S10"))) %>%
+ dplyr::mutate(MFI = as.numeric(MFI)) %>%
+ ggplot(aes(x = Sample_dilution, y = MFI, color = protein, group = protein)) +
+ geom_point() +
+ geom_line() +
+ scale_y_log10(breaks = c(0, 10, 100, 1000, 10000)) +
+ labs(x = "standard curve",
+ y = "log(MFI)") +
+ facet_wrap(~protein) +
+ theme_bw() +
+ theme(axis.text.x = element_text(angle = 90, hjust = 1))
+ }
## Plot plate counts
- plot_counts <- bead_counts %>%
- ggplot(mapping = aes(x = col, y = fct_rev(row), fill = colour), fill = summary)+
- geom_tile(aes(height = 0.90, width = 0.90)) +
- scale_x_continuous(breaks = c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12))+
- scale_fill_manual(values = c("sufficient beads" = "#91bfdb", "repeat" = "#d73027"))+
- theme_linedraw()+
- labs(x = "columns", y = "rows", fill = "")
+ if (nrow(bead_counts) > 0) {
+ plot_counts <- bead_counts %>%
+ ggplot(mapping = aes(x = col, y = fct_rev(row), fill = colour), fill = summary)+
+ geom_tile(aes(height = 0.90, width = 0.90)) +
+ scale_x_continuous(breaks = c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12))+
+ scale_fill_manual(values = c("sufficient beads" = "#91bfdb", "repeat" = "#d73027"))+
+ theme_linedraw()+
+ labs(x = "columns", y = "rows", fill = "")
+ }
## Plot blank sample MFI for each protein - if there is more than one blank sample label as "Blank1", "Blank2" etc
- plot_blank <- blank_MFI %>%
- dplyr::select(-Location) %>%
- dplyr::mutate(Sample = paste0(Sample,1:n())) %>%
- tidyr::pivot_longer(-Sample, names_to = "protein", values_to = "MFI") %>%
- ggplot(aes(x = factor(protein), y = as.numeric(MFI), fill = Sample)) +
- geom_bar(stat = "identity", position = "dodge") +
- geom_hline(yintercept = 50, linetype = "dashed", color = "grey") +
- labs(x = "protein",
- y = "MFI") +
- theme_linedraw() +
- theme(axis.text.x = element_text(angle = 90, hjust = 1))
+ if (nrow(blank_MFI) > 0) {
+ plot_blank <- blank_MFI %>%
+ dplyr::select(-Location) %>%
+ dplyr::mutate(Sample = paste0(Sample,1:n())) %>%
+ tidyr::pivot_longer(-Sample, names_to = "protein", values_to = "MFI") %>%
+ ggplot(aes(x = factor(protein), y = as.numeric(MFI), fill = Sample)) +
+ geom_bar(stat = "identity", position = "dodge") +
+ geom_hline(yintercept = 50, linetype = "dashed", color = "grey") +
+ labs(x = "protein",
+ y = "MFI") +
+ theme_linedraw() +
+ theme(axis.text.x = element_text(angle = 90, hjust = 1))
+ }
## Plot model curves
plots_model <- lapply(seq_along(model_results), function(x){
@@ -526,32 +603,40 @@ getRelativeAntibodyUnits = function(RAW_MFI_FILE_NAME, RAW_MFI_FILE_PATH,
# saveRDS(std_curve, file = paste0(EXP_DIR, "STD_CURVES.rds"))
# Save PDF of std curve plots
- ggsave(plot_stdcurve,
- filename = paste0(EXP_DIR,"STD_CURVES_PLOT.pdf"),
- height = 8,
- width = 12,
- units = "in")
+ if (!is.null(plot_stdcurve)) {
+ ggsave(plot_stdcurve,
+ filename = paste0(EXP_DIR,"STD_CURVES_PLOT.pdf"),
+ height = 8,
+ width = 12,
+ units = "in")
+ }
# Save PDF of bead count plot
- ggsave(plot_counts,
- filename = paste0(EXP_DIR,"PLATE_BEADS_COUNT_PLOT.pdf"),
- height = 8,
- width = 12,
- units = "in")
+ if(!is.null(plot_counts)) {
+ ggsave(plot_counts,
+ filename = paste0(EXP_DIR,"PLATE_BEADS_COUNT_PLOT.pdf"),
+ height = 8,
+ width = 12,
+ units = "in")
+ }
# Save PDF ofblank sample QC plot
- ggsave(plot_blank,
- filename = paste0(EXP_DIR,"BLANK_SAMPLE_PLOT.pdf"),
- height = 4,
- width = 6,
- units = "in")
+ if (!is.null(plot_blank)) {
+ ggsave(plot_blank,
+ filename = paste0(EXP_DIR,"BLANK_SAMPLE_PLOT.pdf"),
+ height = 4,
+ width = 6,
+ units = "in")
+ }
# Save PDF of model plots
- ggsave(all_model_plots,
- filename = paste0(EXP_DIR,"MODEL_PLOTS.pdf"),
- height = 8.27,
- width = 11.69,
- units = "in")
+ if (!is.null(all_model_plots)) {
+ ggsave(all_model_plots,
+ filename = paste0(EXP_DIR,"MODEL_PLOTS.pdf"),
+ height = 8.27,
+ width = 11.69,
+ units = "in")
+ }
## Write to file
write.csv(results.df.wide,
@@ -625,25 +710,25 @@ getAntigenNames = function(RAU_Dilution, ANTIGEN_FILE_PATH, ANTIGEN_CSV){
################# Function 4: Random forest serology analysis
getSeropositiveResults_RF = function(PATHWAY_1,
- RAU_user_id_columns, RAU_user_RAU_columns,
- RAU_RESULTS,
- ANTIGEN_FILE_PATH, ANTIGEN_CSV,
- CHECK_NAME, CHECK_ID, ID, DATE,
- EXP_DIR,
- MODEL_W16,
- MODEL_W16_3_TARGETS,
- MODEL_W16_EQUAL_TARGET,
- MODEL_W16_HIGH_SP_TARGET,
- MODEL_W16_HIGH_SE_TARGET,
- MODEL_W16_OTHER_SE,
- MODEL_W16_OTHER_SP,
- MODEL_W47,
- MODEL_W47_3_TARGETS,
- MODEL_W47_EQUAL_TARGET,
- MODEL_W47_HIGH_SP_TARGET,
- MODEL_W47_HIGH_SE_TARGET,
- MODEL_W47_OTHER_SE,
- MODEL_W47_OTHER_SP){
+ RAU_user_id_columns, RAU_user_RAU_columns,
+ RAU_RESULTS,
+ ANTIGEN_FILE_PATH, ANTIGEN_CSV,
+ CHECK_NAME, CHECK_ID, ID, DATE,
+ EXP_DIR,
+ MODEL_W16,
+ MODEL_W16_3_TARGETS,
+ MODEL_W16_EQUAL_TARGET,
+ MODEL_W16_HIGH_SP_TARGET,
+ MODEL_W16_HIGH_SE_TARGET,
+ MODEL_W16_OTHER_SE,
+ MODEL_W16_OTHER_SP,
+ MODEL_W47,
+ MODEL_W47_3_TARGETS,
+ MODEL_W47_EQUAL_TARGET,
+ MODEL_W47_HIGH_SP_TARGET,
+ MODEL_W47_HIGH_SE_TARGET,
+ MODEL_W47_OTHER_SE,
+ MODEL_W47_OTHER_SP){
##############################################################################################
## ##
@@ -1164,25 +1249,25 @@ getSeropositiveResults_RF = function(PATHWAY_1,
################# Function 5: SVM serology analysis
getSeropositiveResults_SVM = function(PATHWAY_1,
- RAU_user_id_columns, RAU_user_RAU_columns,
- RAU_RESULTS,
- ANTIGEN_FILE_PATH, ANTIGEN_CSV,
- CHECK_NAME, CHECK_ID, ID, DATE,
- EXP_DIR,
- MODEL_W16,
- MODEL_W16_3_TARGETS,
- MODEL_W16_EQUAL_TARGET,
- MODEL_W16_HIGH_SP_TARGET,
- MODEL_W16_HIGH_SE_TARGET,
- MODEL_W16_OTHER_SE,
- MODEL_W16_OTHER_SP,
- MODEL_W47,
- MODEL_W47_3_TARGETS,
- MODEL_W47_EQUAL_TARGET,
- MODEL_W47_HIGH_SP_TARGET,
- MODEL_W47_HIGH_SE_TARGET,
- MODEL_W47_OTHER_SE,
- MODEL_W47_OTHER_SP){
+ RAU_user_id_columns, RAU_user_RAU_columns,
+ RAU_RESULTS,
+ ANTIGEN_FILE_PATH, ANTIGEN_CSV,
+ CHECK_NAME, CHECK_ID, ID, DATE,
+ EXP_DIR,
+ MODEL_W16,
+ MODEL_W16_3_TARGETS,
+ MODEL_W16_EQUAL_TARGET,
+ MODEL_W16_HIGH_SP_TARGET,
+ MODEL_W16_HIGH_SE_TARGET,
+ MODEL_W16_OTHER_SE,
+ MODEL_W16_OTHER_SP,
+ MODEL_W47,
+ MODEL_W47_3_TARGETS,
+ MODEL_W47_EQUAL_TARGET,
+ MODEL_W47_HIGH_SP_TARGET,
+ MODEL_W47_HIGH_SE_TARGET,
+ MODEL_W47_OTHER_SE,
+ MODEL_W47_OTHER_SP){
##############################################################################################
## ##
@@ -1370,30 +1455,30 @@ getSeropositiveResults_SVM = function(PATHWAY_1,
# Create binary seropositivity value
if(PATHWAY_1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:4],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_79SE_79SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative"),
- SEROPOSITIVE_63SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative"),
- SEROPOSITIVE_90SE_59SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_79SE_79SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative"),
+ SEROPOSITIVE_63SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative"),
+ SEROPOSITIVE_90SE_59SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
}else{
if(RAU_user_id_columns == 1){
col_name = colnames(RAU_RESULTS)[1]
SVM_SEROPOS = as.data.frame(cbind(as.character(RAU_RESULTS[,colnames(RAU_RESULTS)[1]]),
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_79SE_79SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative"),
- SEROPOSITIVE_63SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative"),
- SEROPOSITIVE_90SE_59SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_79SE_79SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative"),
+ SEROPOSITIVE_63SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative"),
+ SEROPOSITIVE_90SE_59SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
colnames(SVM_SEROPOS) = c(col_name, colnames(SVM_SEROPOS[2:ncol(SVM_SEROPOS)]))
}
if(RAU_user_id_columns > 1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:as.integer(RAU_user_id_columns)],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_79SE_79SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative"),
- SEROPOSITIVE_63SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative"),
- SEROPOSITIVE_90SE_59SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_79SE_79SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative"),
+ SEROPOSITIVE_63SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative"),
+ SEROPOSITIVE_90SE_59SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
}
}
@@ -1405,30 +1490,30 @@ getSeropositiveResults_SVM = function(PATHWAY_1,
# Create binary seropositivity value
if(PATHWAY_1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:4],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_80SE_80SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative"),
- SEROPOSITIVE_64SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative"),
- SEROPOSITIVE_90SE_60SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_80SE_80SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative"),
+ SEROPOSITIVE_64SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative"),
+ SEROPOSITIVE_90SE_60SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
}else{
if(RAU_user_id_columns == 1){
col_name = colnames(RAU_RESULTS)[1]
SVM_SEROPOS = as.data.frame(cbind(as.character(RAU_RESULTS[,colnames(RAU_RESULTS)[1]]),
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_80SE_80SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative"),
- SEROPOSITIVE_64SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative"),
- SEROPOSITIVE_90SE_60SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_80SE_80SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative"),
+ SEROPOSITIVE_64SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative"),
+ SEROPOSITIVE_90SE_60SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
colnames(SVM_SEROPOS) = c(col_name, colnames(SVM_SEROPOS[2:ncol(SVM_SEROPOS)]))
}
if(RAU_user_id_columns > 1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:as.integer(RAU_user_id_columns)],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_80SE_80SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative"),
- SEROPOSITIVE_64SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative"),
- SEROPOSITIVE_90SE_60SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_80SE_80SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative"),
+ SEROPOSITIVE_64SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative"),
+ SEROPOSITIVE_90SE_60SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
}
}
}
@@ -1443,24 +1528,24 @@ getSeropositiveResults_SVM = function(PATHWAY_1,
# Create binary seropositivity value
if(PATHWAY_1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:4],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_79SE_79SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_79SE_79SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative")), stringsAsFactors = F)
}else{
if(RAU_user_id_columns == 1){
col_name = colnames(RAU_RESULTS)[1]
SVM_SEROPOS = as.data.frame(cbind(as.character(RAU_RESULTS[,colnames(RAU_RESULTS)[1]]),
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_79SE_79SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_79SE_79SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative")), stringsAsFactors = F)
colnames(SVM_SEROPOS) = c(col_name, colnames(SVM_SEROPOS[2:ncol(SVM_SEROPOS)]))
}
if(RAU_user_id_columns > 1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:as.integer(RAU_user_id_columns)],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_79SE_79SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_79SE_79SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative")), stringsAsFactors = F)
}
}
@@ -1472,24 +1557,24 @@ getSeropositiveResults_SVM = function(PATHWAY_1,
# Create binary seropositivity value
if(PATHWAY_1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:4],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_80SE_80SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_80SE_80SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative")), stringsAsFactors = F)
}else{
if(RAU_user_id_columns == 1){
col_name = colnames(RAU_RESULTS)[1]
SVM_SEROPOS = as.data.frame(cbind(as.character(RAU_RESULTS[,colnames(RAU_RESULTS)[1]]),
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_80SE_80SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_80SE_80SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative")), stringsAsFactors = F)
colnames(SVM_SEROPOS) = c(col_name, colnames(SVM_SEROPOS[2:ncol(SVM_SEROPOS)]))
}
if(RAU_user_id_columns > 1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:as.integer(RAU_user_id_columns)],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_80SE_80SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_80SE_80SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_1, "Positive", "Negative")), stringsAsFactors = F)
}
}
}
@@ -1504,24 +1589,24 @@ getSeropositiveResults_SVM = function(PATHWAY_1,
# Create binary seropositivity value
if(PATHWAY_1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:4],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_63SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_63SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative")), stringsAsFactors = F)
}else{
if(RAU_user_id_columns == 1){
col_name = colnames(RAU_RESULTS)[1]
SVM_SEROPOS = as.data.frame(cbind(as.character(RAU_RESULTS[,colnames(RAU_RESULTS)[1]]),
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_63SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_63SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative")), stringsAsFactors = F)
colnames(SVM_SEROPOS) = c(col_name, colnames(SVM_SEROPOS[2:ncol(SVM_SEROPOS)]))
}
if(RAU_user_id_columns > 1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:as.integer(RAU_user_id_columns)],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_63SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_63SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative")), stringsAsFactors = F)
}
}
@@ -1533,24 +1618,24 @@ getSeropositiveResults_SVM = function(PATHWAY_1,
# Create binary seropositivity value
if(PATHWAY_1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:4],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_64SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_64SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative")), stringsAsFactors = F)
}else{
if(RAU_user_id_columns == 1){
col_name = colnames(RAU_RESULTS)[1]
SVM_SEROPOS = as.data.frame(cbind(as.character(RAU_RESULTS[,colnames(RAU_RESULTS)[1]]),
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_64SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_64SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative")), stringsAsFactors = F)
colnames(SVM_SEROPOS) = c(col_name, colnames(SVM_SEROPOS[2:ncol(SVM_SEROPOS)]))
}
if(RAU_user_id_columns > 1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:as.integer(RAU_user_id_columns)],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_64SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_64SE_90SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_2, "Positive", "Negative")), stringsAsFactors = F)
}
}
}
@@ -1565,24 +1650,24 @@ getSeropositiveResults_SVM = function(PATHWAY_1,
# Create binary seropositivity value
if(PATHWAY_1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:4],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_90SE_59SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_90SE_59SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
}else{
if(RAU_user_id_columns == 1){
col_name = colnames(RAU_RESULTS)[1]
SVM_SEROPOS = as.data.frame(cbind(as.character(RAU_RESULTS[,colnames(RAU_RESULTS)[1]]),
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_90SE_59SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_90SE_59SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
colnames(SVM_SEROPOS) = c(col_name, colnames(SVM_SEROPOS[2:ncol(SVM_SEROPOS)]))
}
if(RAU_user_id_columns > 1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:as.integer(RAU_user_id_columns)],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_90SE_59SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_90SE_59SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
}
}
@@ -1594,24 +1679,24 @@ getSeropositiveResults_SVM = function(PATHWAY_1,
# Create binary seropositivity value
if(PATHWAY_1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:4],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_90SE_60SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_90SE_60SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
}else{
if(RAU_user_id_columns == 1){
col_name = colnames(RAU_RESULTS)[1]
SVM_SEROPOS = as.data.frame(cbind(as.character(RAU_RESULTS[,colnames(RAU_RESULTS)[1]]),
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_90SE_60SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_90SE_60SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
colnames(SVM_SEROPOS) = c(col_name, colnames(SVM_SEROPOS[2:ncol(SVM_SEROPOS)]))
}
if(RAU_user_id_columns > 1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:as.integer(RAU_user_id_columns)],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- SEROPOSITIVE_90SE_60SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ SEROPOSITIVE_90SE_60SP=ifelse(SVM_MODEL_VOTES <= CUTOFF_3, "Positive", "Negative")), stringsAsFactors = F)
}
}
}
@@ -1626,27 +1711,27 @@ getSeropositiveResults_SVM = function(PATHWAY_1,
# Create binary seropositivity value
if(PATHWAY_1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:4],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- assign(paste0("SEROPOSITIVE_",as.character(round(as.numeric(SE_Other)*100,digits = 0)), "SE_",as.character(round(as.numeric(SP_Other)*100,digits = 0)),"SP"),ifelse(SVM_MODEL_VOTES <= CUTOFF_4, "Positive", "Negative"))),
- stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ assign(paste0("SEROPOSITIVE_",as.character(round(as.numeric(SE_Other)*100,digits = 0)), "SE_",as.character(round(as.numeric(SP_Other)*100,digits = 0)),"SP"),ifelse(SVM_MODEL_VOTES <= CUTOFF_4, "Positive", "Negative"))),
+ stringsAsFactors = F)
}else{
if(RAU_user_id_columns == 1){
col_name = colnames(RAU_RESULTS)[1]
SVM_SEROPOS = as.data.frame(cbind(as.character(RAU_RESULTS[,colnames(RAU_RESULTS)[1]]),
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- assign(paste0("SEROPOSITIVE_",as.character(round(as.numeric(SE_Other)*100,digits = 0)), "SE_",as.character(round(as.numeric(SP_Other)*100,digits = 0)),"SP"),ifelse(SVM_MODEL_VOTES <= CUTOFF_4, "Positive", "Negative"))),
- stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ assign(paste0("SEROPOSITIVE_",as.character(round(as.numeric(SE_Other)*100,digits = 0)), "SE_",as.character(round(as.numeric(SP_Other)*100,digits = 0)),"SP"),ifelse(SVM_MODEL_VOTES <= CUTOFF_4, "Positive", "Negative"))),
+ stringsAsFactors = F)
colnames(SVM_SEROPOS) = c(col_name, colnames(SVM_SEROPOS[2:ncol(SVM_SEROPOS)]))
}
if(RAU_user_id_columns > 1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:as.integer(RAU_user_id_columns)],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- assign(paste0("SEROPOSITIVE_",as.character(round(as.numeric(SE_Other)*100,digits = 0)), "SE_",as.character(round(as.numeric(SP_Other)*100,digits = 0)),"SP"),ifelse(SVM_MODEL_VOTES <= CUTOFF_4, "Positive", "Negative"))),
- stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ assign(paste0("SEROPOSITIVE_",as.character(round(as.numeric(SE_Other)*100,digits = 0)), "SE_",as.character(round(as.numeric(SP_Other)*100,digits = 0)),"SP"),ifelse(SVM_MODEL_VOTES <= CUTOFF_4, "Positive", "Negative"))),
+ stringsAsFactors = F)
}
}
}
@@ -1657,27 +1742,27 @@ getSeropositiveResults_SVM = function(PATHWAY_1,
# Create binary seropositivity value
if(PATHWAY_1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:4],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- assign(paste0("SEROPOSITIVE_",as.character(round(as.numeric(SE_Other)*100,digits = 0)), "SE_",as.character(round(as.numeric(SP_Other)*100,digits = 0)),"SP"),ifelse(SVM_MODEL_VOTES <= CUTOFF_4, "Positive", "Negative"))),
- stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ assign(paste0("SEROPOSITIVE_",as.character(round(as.numeric(SE_Other)*100,digits = 0)), "SE_",as.character(round(as.numeric(SP_Other)*100,digits = 0)),"SP"),ifelse(SVM_MODEL_VOTES <= CUTOFF_4, "Positive", "Negative"))),
+ stringsAsFactors = F)
}else{
if(RAU_user_id_columns == 1){
col_name = colnames(RAU_RESULTS)[1]
SVM_SEROPOS = as.data.frame(cbind(as.character(RAU_RESULTS[,colnames(RAU_RESULTS)[1]]),
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- assign(paste0("SEROPOSITIVE_",as.character(round(as.numeric(SE_Other)*100,digits = 0)), "SE_",as.character(round(as.numeric(SP_Other)*100,digits = 0)),"SP"),ifelse(SVM_MODEL_VOTES <= CUTOFF_4, "Positive", "Negative"))),
- stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ assign(paste0("SEROPOSITIVE_",as.character(round(as.numeric(SE_Other)*100,digits = 0)), "SE_",as.character(round(as.numeric(SP_Other)*100,digits = 0)),"SP"),ifelse(SVM_MODEL_VOTES <= CUTOFF_4, "Positive", "Negative"))),
+ stringsAsFactors = F)
colnames(SVM_SEROPOS) = c(col_name, colnames(SVM_SEROPOS[2:ncol(SVM_SEROPOS)]))
}
if(RAU_user_id_columns > 1){
SVM_SEROPOS = as.data.frame(cbind(RAU_RESULTS[,1:as.integer(RAU_user_id_columns)],
- exp(RAU_Dilution_Subset),
- SVM_MODEL_VOTES,
- assign(paste0("SEROPOSITIVE_",as.character(round(as.numeric(SE_Other)*100,digits = 0)), "SE_",as.character(round(as.numeric(SP_Other)*100,digits = 0)),"SP"),ifelse(SVM_MODEL_VOTES <= CUTOFF_4, "Positive", "Negative"))),
- stringsAsFactors = F)
+ exp(RAU_Dilution_Subset),
+ SVM_MODEL_VOTES,
+ assign(paste0("SEROPOSITIVE_",as.character(round(as.numeric(SE_Other)*100,digits = 0)), "SE_",as.character(round(as.numeric(SP_Other)*100,digits = 0)),"SP"),ifelse(SVM_MODEL_VOTES <= CUTOFF_4, "Positive", "Negative"))),
+ stringsAsFactors = F)
}
}
}
diff --git a/SHINY_APP.R b/SHINY_APP.R
index 0a3a99c8298c637e6bf3e0bb6bf8ae85ee3240a5..53f5381fa8a2b221c32ee2c73b4cdf169e8a7b71 100644
--- a/SHINY_APP.R
+++ b/SHINY_APP.R
@@ -56,7 +56,7 @@ ui <- fluidPage(
p("The Pv SeroTAT Tool has been developed by Ivo Mueller's research groups at the Institut Pasteur and Walter and Eliza Hall Institut of Medical Research. The tool is a patented diagnostic test based on validated serological markers of recent",tags$em("Plasmodium vivax")," infections and hypnozoite carriage. The tool uses machine learning classification algorithms (Random Forest (RF) and Support Vector Machine (SVM)) to predict sero-positivity with varying performance."),
p("Validation of this tool using Random Forest can be found in the", a("Nature Medicine Publication", href="https://www.nature.com/articles/s41591-020-0841-4")),
p("Please contact on how to cite this tool in publications."),
- p(em("Follow steps 1 - 6 to process Luminex MFI or Relative Antibody Unit (RAU) data and run the diagnostic test.")),
+ p(em("Follow steps 1 - 6 to process Luminex/BioPlex MFI or Relative Antibody Unit (RAU) data and run the diagnostic test.")),
hr()),
tabPanel("Biomarkers of exposure",
br(),
@@ -216,7 +216,7 @@ ui <- fluidPage(
fileInput("MFI_file",
h5("4.1 Load raw MFI file (required)"),
accept = c(".csv", ".xlsx")),
- helpText("Note: a (.csv) or (.xlsx) file of raw outputs of the Luminex-MagPix machine.
+ helpText("Note: a (.csv) or (.xlsx) file of raw outputs of the Luminex-MagPix or BioPlex machine.
If you encounter errors when loading a (.csv) file, convert it to (.xlsx) in Excel and load the (.xlsx) file instead.")),
conditionalPanel("input.radio_data == 1",
textInput("MFI_num_antigens",