cyanoFilter

A Semi-Automated Framework for Identifying Phytplanktons and Cyanobacteria Population in Flow Cytometry Data

Olusoji O. D., Spaak J., Neyens T.,De Laender F., Aerts M.

2024-10-29

Introduction

Flow cytometry is a well-known technique for identifying cell populations contained in a biological smaple. It is largely applied in biomedical and medical sciences for cell sorting, counting, biomarker detections and protein engineering. The technique also provides an energy efficient alternative to microscopy that has long been the standard technique for cell population identification. Cyanobacteria are bacteria phylum believe to contribute more than 50% of atmospheric oxygen via oxygen and are found almost everywhere. These bacteria are also one of the known oldest life forms known to obtain their energy via photosynthesis.

Crucial Synechococcus Properties

Illustrations

We load the package and necessary dependencies below. We also load tidyverse for some data cleaning steps that we need to carry out.

library(dplyr)
library(magrittr)
library(tidyr)
library(purrr)
library(flowCore)
library(flowDensity)
library(cyanoFilter)

To illustrate the funtions contained in this package, we use two datafiles contained by default in the package. These are just demonstration dataset, hence are not documented in the helpfiles.

metadata <- system.file("extdata", "2019-03-25_Rstarted.csv", 
  package = "cyanoFilter", 
  mustWork = TRUE)
metafile <- read.csv(metadata, skip = 7, stringsAsFactors = FALSE, 
  check.names = TRUE)
#columns containing dilution, $\mu l$ and id information
metafile <- metafile %>% 
  dplyr::select(Sample.Number, 
                Sample.ID,
                Number.of.Events,
                Dilution.Factor,
                Original.Volume,
                Cells.L)

Each row in the csv file corresponds to a measurement from two types of cyanobacteria cells carried out at one of three dilution levels. The columns contain information about the dilution level, the number of cells per micro-litre (\(cell/\mu l\)), number of particles measured and a unique identification code for each measurement. The Sample.ID column is structured in the format cyanobacteria_dilution. We extract the cyanobacteria part of this column into a new column and also rename the \(cell/\mu l\) column with the following code:

#extract the part of the Sample.ID that corresponds to BS4 or BS5
metafile <- metafile %>% dplyr::mutate(Sample.ID2 = 
                                         stringr::str_extract(metafile$Sample.ID, "BS*[4-5]")
                                       )
#clean up the Cells.muL column
names(metafile)[which(stringr::str_detect(names(metafile), "Cells."))] <- "CellspML"

Good Measurements

To determine the appropriate data file to read from a FCM datafile, the desired minimum, maximum and column containing the \(cell\mu l\) values are supplied to the goodfcs() function. The code below demonstrates the use of this function for a situation where the desired minimum and maximum for \(cell/\mu l\) is 50 and 1000 respectively.

metafile <- metafile %>% mutate(Status = cyanoFilter::goodFcs(metafile = metafile, 
                                                              col_cpml = "CellspML", 
                                        mxd_cellpML = 1000, 
                                        mnd_cellpML = 50)
                                )
knitr::kable(metafile)
Sample.Number Sample.ID Number.of.Events Dilution.Factor Original.Volume CellspML Sample.ID2 Status
1 BS4_20000 6918 20000 10 62.02270 BS4 good
2 BS4_10000 6591 10000 10 116.76311 BS4 good
3 BS4_2000 6508 2000 10 517.90008 BS4 good
4 BS5_20000 5976 20000 10 48.31036 BS5 bad
5 BS5_10000 5844 10000 10 90.51666 BS5 good
6 BS5_2000 5829 2000 10 400.72498 BS5 good

The function adds an extra column, Status, with entries good or bad to the metafile. Rows containing \(cell/\mu l\) values outside the desired minimum and maximum are labelled bad. Note that the Status column for the fourth row is labelled bad, because it has a \(cell/\mu l\) value outside the desired range.

Files to Retain

Although any of the files labelled good can be read from the FCM file, the retain() function can help select either the file with the highest \(cell/\mu l\) or that with the smallest \(cell/\mu l\) value. To do this, one supplies the function with the status column, \(cell/\mu l\) column and the desired decision. The code below demonstrates this action for a case where we want to select the file with the maximum \(cell/\mu l\) from the good measurements for each unique sample ID.

broken <- metafile %>% group_by(Sample.ID2) %>% nest()
metafile$Retained <- unlist(map(broken$data, function(.x) {
  retain(meta_files = .x, make_decision = "maxi",
  Status = "Status",
  CellspML = "CellspML")
 })
)
knitr::kable(metafile)
Sample.Number Sample.ID Number.of.Events Dilution.Factor Original.Volume CellspML Sample.ID2 Status Retained
1 BS4_20000 6918 20000 10 62.02270 BS4 good No!
2 BS4_10000 6591 10000 10 116.76311 BS4 good No!
3 BS4_2000 6508 2000 10 517.90008 BS4 good Retain
4 BS5_20000 5976 20000 10 48.31036 BS5 bad No!
5 BS5_10000 5844 10000 10 90.51666 BS5 good No!
6 BS5_2000 5829 2000 10 400.72498 BS5 good Retain

This function adds another column, Retained, to the metafile. The third and sixth row in the metadata are with the highest \(cell/\mu l\) values, thus one can proceed to read the fourth and sixth file from the corresponding FCS file for BS4 and BS5 respectively. This implies that we are reading in only two FCS files rather than the six measured files.

Flow Cytometer File Processing

To read B4_18_1.fcs file into R, we use the read.FCS() function from the flowCore package. The dataset option enables the specification of the precise file to be read. Since this datafile contains one file only, we set this option to 1. If this option is set to 2, it gives an error since text.fcs contains only one datafile.

flowfile_path <- system.file("extdata", "B4_18_1.fcs", package = "cyanoFilter",
  mustWork = TRUE)
flowfile <- read.FCS(flowfile_path, alter.names = TRUE,
  transformation = FALSE, emptyValue = FALSE,
  dataset = 1)
flowfile
#> flowFrame object ' B4_18_1'
#> with 8729 cells and 11 observables:
#>            name                   desc     range   minRange  maxRange
#> $P1    FSC.HLin Forward Scatter (FSC..     1e+05    0.00000     99999
#> $P2    SSC.HLin Side Scatter (SSC-HL..     1e+05  -34.47928     99999
#> $P3  GRN.B.HLin Green-B Fluorescence..     1e+05  -21.19454     99999
#> $P4  YEL.B.HLin Yellow-B Fluorescenc..     1e+05  -10.32744     99999
#> $P5  RED.B.HLin Red-B Fluorescence (..     1e+05   -5.34720     99999
#> $P6  NIR.B.HLin Near IR-B Fluorescen..     1e+05   -4.30798     99999
#> $P7  RED.R.HLin Red-R Fluorescence (..     1e+05  -25.49018     99999
#> $P8  NIR.R.HLin Near IR-R Fluorescen..     1e+05  -16.02002     99999
#> $P9    SSC.ALin Side Scatter Area (S..     1e+05    0.00000     99999
#> $P10      SSC.W Side Scatter Width (..     1e+05 -111.00000     99999
#> $P11       TIME                   Time     1e+05    0.00000     99999
#> 368 keywords are stored in the 'description' slot

The R object flowfile contains measurements about 8729 cells across 10 channels since the time channel does not contain any information about the properties of the measured cells.

Transformation and visualisation

To examine the need for transformation, a visual representation of the information in the expression matrix is of great use. The ggpairsDens() function produces a panel plot of all measured channels. Each plot is also smoothed to show the cell density at every part of the plot.

flowfile_nona <- noNA(x = flowfile)
ggpairsDens(flowfile_nona, notToPlot = "TIME")
Panel plot for all channels measured in flowfile_nona. A bivariate kernel smoothed color density is used to indicate the cell density.
Panel plot for all channels measured in flowfile_nona. A bivariate kernel smoothed color density is used to indicate the cell density.

We obtain Figure above by using the ggpairsDens() function after removing all NA values from the expression matrix with the nona() function. There is a version of the function, pairs_plot() that produces standard base scatter plots also smoothed to indicate cell density.


flowfile_noneg <- noNeg(x = flowfile_nona)
flowfile_logtrans <- lnTrans(x = flowfile_noneg, 
  notToTransform = c("SSC.W", "TIME"))
ggpairsDens(flowfile_logtrans, notToPlot = "TIME")
Panel plot for log-transformed channels for flowfile_logtrans. A bivariate kernel smoothed color density is used to indicate the cell density.
Panel plot for log-transformed channels for flowfile_logtrans. A bivariate kernel smoothed color density is used to indicate the cell density.

The second figure is the result of performing a logarithmic transformation in addition to the previous actions taken. The logarithmic transformation appears satisfactory in this case, as it allow a better examination of the information contained in each panel of the figure. Moreover, the clusters are clearly visible in this figure compared to the former figure. Other possible transformation (linear, bi-exponential and arcsinh) can be pursued if the logarithm transformation is not satisfactory. Functions for these transformations are provided in the flowCore package.

Gating

Flow cytometry outcomes can be divided into 3 and they are not entirely mutually exclusive but this is not a problem as scientists are often interested in a pre-defined outcome.

The set of functions below identifies margin events and singlets. Doublets are normally pre-filtered during the event acquiring phase when running the flow cytometer.

The set of functions below identifies margin events and singlets. Doublets are normally pre-filtered during the event

Gating margin events

To remove margin events, the cellmargin() function takes the column in the expression matrix corresponding to measurements about the width of each cell. The code below demonstrates the removal of margin events using the SSC.W column with the option to estimate the cut point between the margin events and the good cells.

flowfile_marginout <- cellMargin(flowframe = flowfile_logtrans,
                                 Channel = 'SSC.W', type = 'estimate', 
                                 y_toplot = "FSC.HLin")
plot(flowfile_marginout)
Smoothed Scatterplot of measured width (SSC.W) and height (FSC.HLin). The red line is the estimated cut point by flowDensity, and every particle below the red line has their width properly measured.
Smoothed Scatterplot of measured width (SSC.W) and height (FSC.HLin). The red line is the estimated cut point by flowDensity, and every particle below the red line has their width properly measured.

summary(flowfile_marginout, 
       channels = c('FSC.HLin', 'SSC.HLin', 
                    'SSC.W'))
#>       Length        Class         Mode 
#>            1 MarginEvents           S4

flowfile_marginout is an S4 object of class MarginEvents with summary(), plot(), fullFlowframe() and reducedFlowframe() methods. Running plot() on flowfile_marginout produces a plot of the width channel against the channel supplied in y_toplot. This action returns the figure @ref(fig:marginEvents). flowfile_marginout contains the following slots:

Running plot() on flowfile_marginout gives you the number of margin and non-margin particles as well as descriptives on channels supplied. These descriptives are computed on the flowfile after the margin events have been removed.

Gating Debris

To identify debris, we leverage on the presence of chlorophyll a


cells_nodebris <-  debrisNc(flowframe = reducedFlowframe(flowfile_marginout), 
                             ch_chlorophyll = "RED.B.HLin", ch_p2 = "YEL.B.HLin",
                             ph = 0.05)
plot(cells_nodebris)
Smoothed Scatterplot of measured chlorophyll a channel (RED.B.HLin) and phycoerythrin channel (YEL.B.HLin). The red lines are the estimated minimum intersection points between the detected peaks.
Smoothed Scatterplot of measured chlorophyll a channel (RED.B.HLin) and phycoerythrin channel (YEL.B.HLin). The red lines are the estimated minimum intersection points between the detected peaks.

Gating cyanobacteria

The phyto_filter() function employs the following algorithm to separate particles into different clusters;

  1. Search for peaks along the supplied pigment and cell complexity channels.
  2. Idneify the minimum intersection point between the peaks observed these channels.
  3. Divide particles into groups based on the minimum intersection points identified in 1 and label each group.
  4. Formulate all possible combinations of labels in step 2.
  5. Assign a new label to the combinations in 3.
  6. Retain clusters that make up a desired proportion of the total number of particles clustered.

bs4_gate1 <- phytoFilter(flowfile = reducedFlowframe(cells_nodebris),
               pig_channels = c("RED.B.HLin", "YEL.B.HLin", "RED.R.HLin"),
               com_channels = c("FSC.HLin", "SSC.HLin"))

plot(bs4_gate1)
Smoothed Scatterplot of all channels used in the gating process.
Smoothed Scatterplot of all channels used in the gating process.

The resulting object is a S4 object of class PhytoFilter with the following slots:

Acknowledgements