qb_from_fits: Extract Q, B and the intrinsic CV0 from fitting results

In flowQB: Automated Quadratic Characterization of Flow Cytometer Instrument Sensitivity: Q, B and CV instrinsic calculations

Description

This function can be used to compute (1) flow cytometer's detection efficiency (i.e., Q, the statistical number of single photoelectrons (Spe) generated per unit of dye in the sample), (2) background illumination (i.e., B, the background light in dye equivalents that sets the minimum variance that underlies all measurements), and (3) the intrinsic CV0 (i.e, variance in the signal produced by the variation in dye amount of beads containing a "fixed" level of dye plus the illumination variations due to particles taking different flow paths through the laser beam) from fitting results produced by either LED data fitting (fit_led function) or bead data fitting (fit_beads, fit_spherotech or fit_thermo_fisher functions). One can calculate based on any of the fits (i.e, the fits, dye_fits, iterated_fits or iterated_dye_fits items from the result list).

Usage

    qb_from_fits(fits)

Arguments

fits

Fitting results as produced by either LED data fitting (fit_led function) or bead data fitting (fit_beads, fit_spherotech or fit_thermo_fisher functions). One can calculate based on any of the fits provided as part of the results of these functions, namely the fits, dye_fits, iterated_fits or iterated_dye_fits items from the result list.

Details

As explained in our paper, this method calculats QI as 1/c1, BSpe = c0/c1^2 and CV0^2 = c2 from the results of quadratic fit for the measured means and variances to a statistical model involving the Poisson distribution expectations for the relation between them. For fitting results containing both, a quadratic and a linear fit coeficients, (i.e, results of LED fitting), this method also includes QI and BSpe from the linear model.

Value

The value is a matrix with columns corresponding to the columns of the input data frame (i.e., the names of the dyes). The rows are as follows: q_QI as the QI from the fits of the quadratic model, q_BSpe as the BSpe from the fits of the quadratic model, q_CV0sq as the CV0^2 from the fits of the quadratic model, l_QI as the QI from the fits of the linear model (provided only if linear fit coefficients are also present in the input data frame, i.e., for LED fits only), and l_BSpe as the BSpe from the fits of the linear model (again, provided only if linear fit coefficients are also present in the input data frame, i.e., for LED fits only)

Author(s)

Josef Spidlen, Wayne Moore, Faysal El Khettabi

See Also

fit_led, fit_beads, fit_spherotech, fit_thermo_fisher

Examples

## Example is based on LED data from the flowQBData package
fcs_directory <- system.file("extdata", "SSFF_LSRII", "LED_Series", 
    package="flowQBData")
fcs_file_path_list <- list.files(fcs_directory, "*.fcs", full.names= TRUE)
## We are working with these FCS files:
basename(fcs_file_path_list)

## Various house keeping information
## - Which channels should be ignored, typically the non-fluorescence
##    channels, such as the time and the scatter channels
ignore_channels <- c("Time", 
    "FSC-A", "FSC-W", "FSC-H", 
    "SSC-A", "SSC-W", "SSC-H")
## - Which dyes would you typically use with the detectors
dyes <- c("APC", "APC-Cy7", "APC-H7", "FITC", "PE", "PE-Cy7", "PerCP", 
    "PerCP-Cy55", "V450", "V500-C")
## - What are the corresponding detectors, provide a vector of short channel 
## names, i.e., values of the $PnN FCS keywords.
detectors <- c("APC-A", "APC-Cy7-A", "APC-Cy7-A", "FITC-A", "PE-A", "PE-Cy7-A",
    "PerCP-Cy5-5-A", "PerCP-Cy5-5-A", "Pacific Blue-A", "Aqua Amine-A")
## - The signal type that you are looking at (Area or Height)
signal_type <- "Area"
## - The instrument make/model
instrument_name <- 'LSRII'
## - Set the minimum and maximum values, peaks with mean outsize of this range
## will be ignored
bounds <- list(minimum = -100, maximum = 100000)
## - The minimum number of usable peaks (represented by different FCS files
## in case of an LED pulser) required in order for a fluorescence channel 
## to be included in the fitting. Peaks with mean expression outside of the 
## bounds specified above are omitted and therefore not considered useful
minimum_fcs_files <- 3 # The default 3 seems to be work well in typical cases
## - What is the maximum number of iterations for iterative fitting with
## weight adjustments
max_iterations <- 10 # The default 10 seems to be enough in typical cases

## Now, let's calculate the fitting
led_results <- fit_led(fcs_file_path_list, ignore_channels, dyes,
    detectors, signal_type, instrument_name, bounds = bounds,
    minimum_useful_peaks = minimum_fcs_files, max_iterations = max_iterations)

qb_from_fits(led_results$iterated_dye_fits)

flowQB documentation built on May 6, 2019, 3:05 a.m.