tcpl 2.0 Data Retrieval

Data Retrieval

This vignette describes how the user can retrieve data from the csv files using tcplLite or from the ToxCast databases using tcpl. Listed below are case examples that extract summary information about the chemical, Bisphenol A, and the assay, NVS_NR_hTRa. The multiple concentration table for level 5 can be downloaded from the publicly available ToxCast data https://www.epa.gov/chemical-research/exploring-toxcast-data-downloadable-data.

A. Using TcplLite

In this example, we will identify the positive responses for the chemical Bisphenol A.

## Load mc tables from levels of interest
mc4 <- tcplLoadData(lvl = 4, type = "mc")
mc5 <- tcplLoadData(lvl = 5, type = "mc")
mc6 <- tcplLoadData(lvl = 6, type = "mc")

## Find all sample IDs (spids) associated with Bisphenol A
chnm <- 'Bisphenol A'
ch <- fread(system.file("/csv/chemical.csv",
                   package = "tcpl"),
                   sep = ",",
                   header = TRUE)
chem <- tcplLoadChem('chnm', chnm)
dat4spid <- mc4[mc4$spid %in% chem$spid]

Now that we isolated all the sample IDs (spids) for the chemical of interest, we can start mapping data from the multiple-concentration tables: level 4(mc4), level 5(mc5), and level 6(mc6) from the local csv directory. Level 4 includes the spids, and level 5 specifies whether the chemical is active or inactive. Level 6 indicates flags reflecting the quality of the data or the analysis.

Next, we match the the m5ids that correspond to the spids of interest. From those, we select the positive responses, with hit call = 1.

dat5spid <- mc5[mc5$m4id %in% dat4spid$m4id]
dat5hit <- dat5spid[hitc == 1]

To identify the flags for the Bisphenol A samples, we match the flag columns in the mc6 table by the corresponding m5ids.

mc6_flags <- mc6[ , .( flag = paste(flag, collapse=";")),
                    by = m5id]
dat5dat6 <- merge(x = mc6_flags,
                  y = dat5hit,
                  by = "m5id", all.y = TRUE)

B. Using Tcpl

We illustrate here the necessary steps for extracting information about Bisphenol A using tcpl rather than tcplLite.

First, we change the driver in tcpl settings to MySQL, and the database to invitrodb :

tcplConf(
          user = 'XXXXXXX', 
          pass = 'XXXXXXX', 
          host = 'XXXXXXX', 
          db = 'invitrodb', 
          drvr = 'MySQL')

Next, we define the chemical of interest and load the related spids:

chnm <- 'Bisphenol A'
chem <- tcplLoadChem('chnm', chnm)

Then, we load the data from different levels to summarize positive responses for this chemical.

dat5 <- tcplPrepOtpt(tcplLoadData
                     (lvl = 5, fld = 'spid', 
                       val = chem$spid, type = 'mc'))
## For positives only, hit call (hitc) should equal 1
dat5 <- dat5[hitc == 1] 
dat6 <- tcplPrepOtpt(tcplLoadData
                     (lvl = 6, fld = 'spid', val = chem$spid, 
                       type = 'mc'))
mc6_mthds <- dat6[ , .( mc6_mthd_id = 
                          paste(mc6_mthd_id, collapse = ",")), 
                   by = m4id]
mc6_flags <- dat6[ , .( flag = 
                          paste(flag, collapse = ";")), 
                   by = m4id]

Then, we can generate all level 6 plots for positive responses for this chemical:

m4ids <- dat5[ , m4id]
graphics.off()
pdf(file = file.path(getwd(),
                   paste("mc6",
                         paste(chnm,collapse = "."),
                         format(Sys.Date(),
                                "%y%m%d.pdf"),
                         sep = "_")),
    height = 6,
    width = 10,
    pointsize = 10)
tcplPlotM4ID(m4ids, lvl = 6)
graphics.off()

In the following example, we will obtain summary information about the example assay NVS_NR_hTRa_Antagonist using tcpl:

## List the assay source IDs
tcplLoadAsid() 
## Find the assay source (NVS)
nvs.assays <- tcplLoadAeid(fld='asid', val = 5) 
## Find the assay name (hTRa)
aeids <- nvs.assays[grep("hTRa", aenm)]$aeid 
## Load the mc5 to determine hit call and summary information
dat5 <- tcplPrepOtpt(tcplLoadData(lvl = 5, type = 'mc', 
                                  fld = 'aeid', val = aeids)) 
dat5 <- dat5[hitc == 1]
## Make the level 6 plots for the positive responses
assay <- 'NVS_NR_hTRa_Antagonist'
m4ids <- dat5[ , m4id]
graphics.off()
pdf(file = file.path(getwd(),
                   paste("mc6",
                         paste(assay,collapse = "."),
                         format(Sys.Date(),
                                "%y%m%d.pdf"),
                         sep = "_")),
    height = 6,
    width = 10,
    pointsize = 10)
tcplPlotM4ID(m4ids, lvl = 6)
graphics.off()

Moreover, we can extract the mc3 data used for plots:

spids <- unique(dat5[,spid])
## logc = log10concentration, starting with micromolar
## units (x-axis), resp = normalized response value (y-axis)
mc3 <- tcplPrepOtpt(tcplLoadData
                    (lvl = 3, type = 'mc', fld = 'spid',
                      val = spids))
mc3 <- mc3[aeid %in% aeids]

Then, we can visualize the normalized mc3 data without tcpl curve-fitting:

library(ggplot2)

graphics.off()
pdf(file = file.path(getwd(),
                   paste("mc3",
                         paste(assay,collapse = "."),
                         format(Sys.Date(),
                                "%y%m%d.pdf"),
                         sep = "_")),
    height = 6,
    width = 10,
    pointsize = 10)

by(mc3, mc3$spid, function (x){
  ggplot(x, aes(x = logc, y = resp), tab) +
    geom_point(aes(group=spid, color=spid)) +
    theme(axis.ticks.x = element_blank(),
          axis.text.x = element_text(size=rel(0.5)),
          axis.ticks.y = element_blank(),
          axis.text.y = element_blank()) +
    xlab("logc") + ylab("resp")
}
)
graphics.off()

The tcplLoadData function can be used to load data from level 7. The added level allows for estimating the uncertainty in the fitted parameters, such as AC₅₀.