In BioData/LabguruR: Labguru API wrapper for R
knitr::opts_chunk$set(
collapse = TRUE,
comment = "#>",
eval = FALSE
)
Introduction to the ic50 package
library(ic50)
Functions within the ic50 package.
- ic50() to launch the GUI
- hts.96() to analyse a plate with 96 wells
- hts.384() to analyse a a plate with 384 wells
ic50()
Example of running the 384 wells analysis with local files.
dir <- "../data-raw/ic50/experiment-001/"
r <- hts.384(indir = paste0(dir, "Files"),
plates = 2,
measure = paste0(dir, "Suppl-files/mpi384_measure.txt"),
control = paste0(dir, "Suppl-files/mpi384_control.txt"),
dilution = paste0(dir, "Suppl-files/mpi384_dilution.txt"),
inhib = rep(0.5,7),
outdir = paste0(dir, "Output"),
normalize = "mean",
graphics = "single")
class(r)
r
Downloading plates data from a Labguru server
Setting up Labguru connection
# Load LabgururR locally
devtools::load_all()
# Authenticate
email <- ""
password <- ""
labguru_authenticate(email, password)
Download JSON data
Loading JSON for plate 271
server <- Sys.getenv("LABGURU_SERVER")
token <- Sys.getenv("LABGURU_TOKEN")
# "https://jonathan.labguru.com/api/v1/plates/271"
base_url <- server
path <- "/api/v1/plates/271"
query <- paste0("token=", token)
url <- httr::modify_url(url = base_url,
path = path,
query = query)
resp <- httr::GET(url)
# Expect resp to be JSON
if (httr::http_type(resp) != "application/json") {
stop("API did not return JSON", call. = FALSE)
}
# Parse without simplifaction for consistency
parsed <- jsonlite::fromJSON(httr::content(resp, as = "text"),
simplifyVector = FALSE,
simplifyDataFrame = TRUE,
flatten = TRUE)
# check for request error
if (httr::http_error(resp)) {
stop(sprintf("API request failed [%s]\n%s", parsed$status, parsed$error), call. = FALSE)
}
Explore data
Further parse the JSON deom the data element
prs_data <- jsonlite::fromJSON(parsed$data)
The parsed data has several elements:
names(prs_data)
- 'wells' contains the measure, dilution and control data
- 'samples' contains information on the cell lines and compounds
- 'added_samples' contains the compound id's
- 'arrange_in' is unknown information
- 'properties' is unknown information
An overview of first 10 columns of data stored in 'samples'. In the first two columns are the id's and names of the compounds.
prs_data$samples[1:10]
Information about the individual wells is stored in several data frames under prs_data\$wells\$samples_metadata. Each data frame contains the wells information for 1 compound and is named after the id of that compound.
names(prs_data$wells$samples_metadata)
A sample of compound 41, which we know from the samples data represents '17AAG'.
head(prs_data$wells$samples_metadata["41"])
Prepare data for ic50
Reorganise samples_metadata.
Assumptions:
- prs_data\$wells\$coordinates is in the same order as the samples_metadata data frames, hence we can column bind them.
- if samples_metadata concentration data is NA the other columns also do not hold valuable information, hence we can filter by it.
- if we filter concentration for NA all stocks data left is numeric
- if the serial_factor column is there then it is always a data frame with a value and symbol column with equal length.
# PREPARE each compound data frame separately
dfs <- lapply(X = prs_data$wells$samples_metadata,
FUN = function(d) {
# Add coordinates before filter
d$coordinates <- prs_data$wells$coordinates
# Serial_factor is a nested dataframe, unnest it
if ("serial_factor" %in% names(d)) {
d$comp_serial_factor_value <- d$serial_factor$value
d$comp_serial_factor_symbol <- d$serial_factor$symbol
d$serial_factor <- NULL
}
# Filter empty rows
d <- d[!is.na(d$concentration), ]
# as.numeric stocks now that only numreic data remains
d$stocks <- as.numeric(d$stocks)
d
})
# SUBSET the compound columns
# ASSUMPTION: The column prs_data$samples$collection_name will be "Compounds" if it is a compound. Can't be missing or spelt differently
collection_names <- prs_data$samples$collection_name
names(collection_names) <- prs_data$samples$id
nc <- collection_names[names(dfs)] == "Compounds"
# BIND ROWS
comp <- Reduce(f = rbind,
x = dfs[nc])
names(comp)[1:2] <- paste("comp", names(comp)[1:2], sep = "_")
# ASSUMPTION: THERE CAN ONLY BE ONE CELL LINE
df <- merge(dfs[!nc][[1]], comp, by.x = "coordinates", by.y = "coordinates")
# TEST: CELL LINE AND COMPOUND SAME LENGTH
nrow(df) == nrow(dfs[!nc][[1]])
nrow(df) == nrow(comp)
# TEST: NO DUPLICATES IN COORDINATES
length(unique(df$coordinates)) == nrow(df)
# ADD: compound names
compound_names <- prs_data$samples$name
names(compound_names) <- as.character(prs_data$samples$id)
df$comp_name <- unname(compound_names[as.character(df$comp_stocks)])
# ADD row and col for printing
df$row <- vapply(substr(df$coordinates, 1, 1), function(c) { which(c == LETTERS) }, FUN.VALUE = numeric(1))
df$col <- as.numeric(gsub("[^0-9.-]+", "", df$coordinates))
# ARRANGE: by row and col
df <- df[order( df[,"row"], df[,"col"] ),]
head(df, 20)
Write measure file
# ASSUMPTION: 1 compound per row
# CHECK: does it require equal columns per row
measure <- lapply(X = unique(df$row),
FUN = function(r) {
d <- df[df$row == r, ]
name <- paste0('"', unique(d$comp_name)[1], '"')
coords <- paste0('"', d$row, ',', d$col, '"')
c(name, coords)
})
names(measure) <- paste0("row", seq_along(measure))
measure <- as.data.frame(t(as.data.frame(measure)))
measure
write.table(x = measure,
file = "measure_from_api.txt",
append = FALSE,
sep = "\t",
row.names = FALSE,
col.names = FALSE,
quote = FALSE)
Write dilution file
# ASSUMPTION: 1 compound per row
# CHECK: does it require equal columns per row
dilution <- lapply(X = unique(df$row),
FUN = function(r) {
d <- df[df$row == r, ]
name <- paste0('"', unique(d$comp_name)[1], '"')
coords <- paste0('"', d$comp_concentration, '"')
c(name, coords)
})
names(dilution) <- paste0("row", seq_along(dilution))
dilution <- as.data.frame(t(as.data.frame(dilution)))
dilution
write.table(x = dilution,
file = "dilution_from_api.txt",
append = FALSE,
sep = "\t",
row.names = FALSE,
col.names = FALSE,
quote = FALSE)
Write control file
head(cntrl, 6)
cntrl <- prs_data$wells[c("control", "coordinates")]
cntrl <- cntrl[cntrl$control, ]
# ADD row and col for printing
cntrl$row <- vapply(substr(cntrl$coordinates, 1, 1), function(c) { which(c == LETTERS) }, FUN.VALUE = numeric(1))
cntrl$col <- as.numeric(gsub("[^0-9.-]+", "", cntrl$coordinates))
# ADD compound name
# ASSUMTION: each row has only 1 compound name
temp <- unique(df[c("comp_name", "row")])
compound_names <- temp$comp_name
names(compound_names) <- as.character(temp$row)
cntrl$comp_name <- unname(compound_names[as.character(cntrl$row)])
# ASSUMPTION: 1 compound per row
# CHECK: does it require equal columns per row
control <- lapply(X = unique(cntrl$row),
FUN = function(r) {
d <- cntrl[cntrl$row == r, ]
name <- paste0('"', unique(d$comp_name)[1], '"')
coords <- paste0('"', d$row, ',', d$col, '"')
c(name, coords)
})
names(control) <- paste0("row", seq_along(control))
control <- as.data.frame(t(as.data.frame(control)))
control
write.table(x = control,
file = "control_from_api.txt",
append = FALSE,
sep = "\t",
row.names = FALSE,
col.names = FALSE,
quote = FALSE)
ic50 analysis
Store data
Store images
pdf_img <- magick::image_read_pdf("dose_response_curves.pdf")
magick::image_info(pdf_img)
magick::image_write(pdf_img[1], "test.png")
OLD STUFF
# comp_id <- names(prs_data$wells$samples_metadata)
# df <- mapply(prs_data$wells$samples_metadata,
# comp_id,
# FUN = function(d, c) {
# names(d) <- paste(names(d), c, sep = "_")
# d
# })
# df <- Reduce(cbind, df)
# View(df)
head(prs_data$wells$samples_metadata[["41"]], 30)
head(prs_data$wells$samples_metadata[["41"]], 30)
names(prs_data)
names(prs_data$wells)
names(prs_data$samples)
prs_data[3:6]
prs_data$samples
head(prs_data$wells)
str(prs_data)
substr(parsed$data, 1, 3000)
substr(parsed$data, 1, 1000)
BioData/LabguruR documentation built on May 11, 2022, 11:46 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
knitr::opts_chunk$set( collapse = TRUE, comment = "#>", eval = FALSE )
Introduction to the ic50 package
library(ic50)
Functions within the ic50 package.
- ic50() to launch the GUI
- hts.96() to analyse a plate with 96 wells
- hts.384() to analyse a a plate with 384 wells
ic50()
Example of running the 384 wells analysis with local files.
dir <- "../data-raw/ic50/experiment-001/" r <- hts.384(indir = paste0(dir, "Files"), plates = 2, measure = paste0(dir, "Suppl-files/mpi384_measure.txt"), control = paste0(dir, "Suppl-files/mpi384_control.txt"), dilution = paste0(dir, "Suppl-files/mpi384_dilution.txt"), inhib = rep(0.5,7), outdir = paste0(dir, "Output"), normalize = "mean", graphics = "single") class(r) r
Downloading plates data from a Labguru server
Setting up Labguru connection
# Load LabgururR locally devtools::load_all() # Authenticate email <- "" password <- "" labguru_authenticate(email, password)
Download JSON data
Loading JSON for plate 271
server <- Sys.getenv("LABGURU_SERVER") token <- Sys.getenv("LABGURU_TOKEN") # "https://jonathan.labguru.com/api/v1/plates/271" base_url <- server path <- "/api/v1/plates/271" query <- paste0("token=", token) url <- httr::modify_url(url = base_url, path = path, query = query) resp <- httr::GET(url) # Expect resp to be JSON if (httr::http_type(resp) != "application/json") { stop("API did not return JSON", call. = FALSE) } # Parse without simplifaction for consistency parsed <- jsonlite::fromJSON(httr::content(resp, as = "text"), simplifyVector = FALSE, simplifyDataFrame = TRUE, flatten = TRUE) # check for request error if (httr::http_error(resp)) { stop(sprintf("API request failed [%s]\n%s", parsed$status, parsed$error), call. = FALSE) }
Explore data
Further parse the JSON deom the data element
prs_data <- jsonlite::fromJSON(parsed$data)
The parsed data has several elements:
names(prs_data)
- 'wells' contains the measure, dilution and control data
- 'samples' contains information on the cell lines and compounds
- 'added_samples' contains the compound id's
- 'arrange_in' is unknown information
- 'properties' is unknown information
An overview of first 10 columns of data stored in 'samples'. In the first two columns are the id's and names of the compounds.
prs_data$samples[1:10]
Information about the individual wells is stored in several data frames under prs_data\$wells\$samples_metadata. Each data frame contains the wells information for 1 compound and is named after the id of that compound.
names(prs_data$wells$samples_metadata)
A sample of compound 41, which we know from the samples data represents '17AAG'.
head(prs_data$wells$samples_metadata["41"])
Prepare data for ic50
Reorganise samples_metadata.
Assumptions:
- prs_data\$wells\$coordinates is in the same order as the samples_metadata data frames, hence we can column bind them.
- if samples_metadata concentration data is NA the other columns also do not hold valuable information, hence we can filter by it.
- if we filter concentration for NA all stocks data left is numeric
- if the serial_factor column is there then it is always a data frame with a value and symbol column with equal length.
# PREPARE each compound data frame separately dfs <- lapply(X = prs_data$wells$samples_metadata, FUN = function(d) { # Add coordinates before filter d$coordinates <- prs_data$wells$coordinates # Serial_factor is a nested dataframe, unnest it if ("serial_factor" %in% names(d)) { d$comp_serial_factor_value <- d$serial_factor$value d$comp_serial_factor_symbol <- d$serial_factor$symbol d$serial_factor <- NULL } # Filter empty rows d <- d[!is.na(d$concentration), ] # as.numeric stocks now that only numreic data remains d$stocks <- as.numeric(d$stocks) d })
# SUBSET the compound columns # ASSUMPTION: The column prs_data$samples$collection_name will be "Compounds" if it is a compound. Can't be missing or spelt differently collection_names <- prs_data$samples$collection_name names(collection_names) <- prs_data$samples$id nc <- collection_names[names(dfs)] == "Compounds"
# BIND ROWS comp <- Reduce(f = rbind, x = dfs[nc]) names(comp)[1:2] <- paste("comp", names(comp)[1:2], sep = "_")
# ASSUMPTION: THERE CAN ONLY BE ONE CELL LINE df <- merge(dfs[!nc][[1]], comp, by.x = "coordinates", by.y = "coordinates")
# TEST: CELL LINE AND COMPOUND SAME LENGTH nrow(df) == nrow(dfs[!nc][[1]]) nrow(df) == nrow(comp) # TEST: NO DUPLICATES IN COORDINATES length(unique(df$coordinates)) == nrow(df)
# ADD: compound names compound_names <- prs_data$samples$name names(compound_names) <- as.character(prs_data$samples$id) df$comp_name <- unname(compound_names[as.character(df$comp_stocks)])
# ADD row and col for printing df$row <- vapply(substr(df$coordinates, 1, 1), function(c) { which(c == LETTERS) }, FUN.VALUE = numeric(1)) df$col <- as.numeric(gsub("[^0-9.-]+", "", df$coordinates))
# ARRANGE: by row and col df <- df[order( df[,"row"], df[,"col"] ),] head(df, 20)
Write measure file
# ASSUMPTION: 1 compound per row # CHECK: does it require equal columns per row measure <- lapply(X = unique(df$row), FUN = function(r) { d <- df[df$row == r, ] name <- paste0('"', unique(d$comp_name)[1], '"') coords <- paste0('"', d$row, ',', d$col, '"') c(name, coords) }) names(measure) <- paste0("row", seq_along(measure)) measure <- as.data.frame(t(as.data.frame(measure))) measure
write.table(x = measure, file = "measure_from_api.txt", append = FALSE, sep = "\t", row.names = FALSE, col.names = FALSE, quote = FALSE)
Write dilution file
# ASSUMPTION: 1 compound per row # CHECK: does it require equal columns per row dilution <- lapply(X = unique(df$row), FUN = function(r) { d <- df[df$row == r, ] name <- paste0('"', unique(d$comp_name)[1], '"') coords <- paste0('"', d$comp_concentration, '"') c(name, coords) }) names(dilution) <- paste0("row", seq_along(dilution)) dilution <- as.data.frame(t(as.data.frame(dilution))) dilution
write.table(x = dilution, file = "dilution_from_api.txt", append = FALSE, sep = "\t", row.names = FALSE, col.names = FALSE, quote = FALSE)
Write control file
head(cntrl, 6)
cntrl <- prs_data$wells[c("control", "coordinates")] cntrl <- cntrl[cntrl$control, ] # ADD row and col for printing cntrl$row <- vapply(substr(cntrl$coordinates, 1, 1), function(c) { which(c == LETTERS) }, FUN.VALUE = numeric(1)) cntrl$col <- as.numeric(gsub("[^0-9.-]+", "", cntrl$coordinates)) # ADD compound name # ASSUMTION: each row has only 1 compound name temp <- unique(df[c("comp_name", "row")]) compound_names <- temp$comp_name names(compound_names) <- as.character(temp$row) cntrl$comp_name <- unname(compound_names[as.character(cntrl$row)]) # ASSUMPTION: 1 compound per row # CHECK: does it require equal columns per row control <- lapply(X = unique(cntrl$row), FUN = function(r) { d <- cntrl[cntrl$row == r, ] name <- paste0('"', unique(d$comp_name)[1], '"') coords <- paste0('"', d$row, ',', d$col, '"') c(name, coords) }) names(control) <- paste0("row", seq_along(control)) control <- as.data.frame(t(as.data.frame(control))) control
write.table(x = control, file = "control_from_api.txt", append = FALSE, sep = "\t", row.names = FALSE, col.names = FALSE, quote = FALSE)
ic50 analysis
Store data
Store images
pdf_img <- magick::image_read_pdf("dose_response_curves.pdf") magick::image_info(pdf_img) magick::image_write(pdf_img[1], "test.png")
OLD STUFF
# comp_id <- names(prs_data$wells$samples_metadata)
# df <- mapply(prs_data$wells$samples_metadata, # comp_id, # FUN = function(d, c) { # names(d) <- paste(names(d), c, sep = "_") # d # }) # df <- Reduce(cbind, df) # View(df)
head(prs_data$wells$samples_metadata[["41"]], 30)
head(prs_data$wells$samples_metadata[["41"]], 30)
names(prs_data)
names(prs_data$wells) names(prs_data$samples)
prs_data[3:6]
prs_data$samples
head(prs_data$wells)
str(prs_data)
substr(parsed$data, 1, 3000)
substr(parsed$data, 1, 1000)
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.