I'm trying to incorporate tmixFilter into a flowCore/flowWorkspace-based FACS data analysis. The primary problem I'm having is that I need to select multiple clusters the tmixFilter results and I can't find any examples a "flowCore"-y way of doing that.

Below is what I have been doing, but it's a complicated workflow that doesn't fit in well with the flowCore-family of workflows. For example, I'd like to incorporate this gating into a GatingWorkflow and be able to use the various plotting functions to visualize gating and backgating.

# 'fs' is a flowSet object
ssca_fsca_tmix <- flowCore::filter(fs, tmixFilter('SSCA_FSCA', c("FSC-A","SSC-A"), K=6, level=0.95))

# now I need to figure out which of the 6 clusters per flowFrame are the ones I need
# there are usually two
FSC_A.bounds <- c(70000,170000)
SSC_A.bounds <- c(10000,50000)

tmix_gate_summary <- tibble(i=seq_along(ssca_fsca_tmix)) %>% 
    # pull out the raw FSC-A and SSC-A data and combine it with the tmixFilter cluster assignment
    mutate(raw.data = map(i, ~tibble(
        `FSC-A`=fs[[.x]]@exprs[,"FSC-A"],
        `SSC-A`=fs[[.x]]@exprs[,"SSC-A"],
        cluster=ssca_fsca_tmix@.Data[[.x]]@subSet
    ))) %>%
    # calculate median FSC-A and SSC-A values for each tmixFilter cluster
    mutate(summary.data = map(raw.data, function(x){
        x %>% dplyr::filter(!is.na(cluster)) %>% 
            group_by(cluster) %>%
            summarize(`FSC-A`=median(`FSC-A`),`SSC-A`=median(`SSC-A`))
    })) %>%
    # identify the clusters with medians in the desired ranges
    mutate(selected.clusters = map(summary.data, function(x){
        x %>% dplyr::filter(
                between(`FSC-A`, FSC_A.bounds[1], FSC_A.bounds[2]),
                between(`SSC-A`, SSC_A.bounds[1], SSC_A.bounds[2])
            ) %>% pull(cluster)
    }))

# use flowClust::split to pull out the desired clusters from each flowFrame
flowFrames <- lapply(seq_along(fs), function(j) {
    cluster_numbers <- tmix_gate_summary %>% 
        dplyr::filter(i == j) %>% 
        pull(selected.clusters) %>%
        unlist
    split(fs[[j]], ssca_fsca_tmix[[j]], population=list(my_filter=cluster_numbers))
})

# combine these back into a new flowSet
fs_gated <- flowSet(lapply(flowFrames, function(x) x$my_filter ))

As an alternate workflow, I did come across cytoUtils::tmix2DGate. I can get the clusters out of that, and pull them out at as ellipse gates, but I can't figure out how to create gates which are the union of these multiple ellipse gates per flowFrame.

flowCore::workflow was deprecated by flowWorkspace::GatingSet for representing/storing/manipulating the gating trees and gated subpopulation data. Take a look at the vignette of flowWorkspace for the basics of adding/querying the gates through GatingSet. On top of that, you can further automate your gating process by using openCyto. Regarding to the union of gates, it can achieved by flowWorkspace::booleanFilter or (boolGate method within the context of openCyto)