stats.R

## Warning: replacing previous import 'ncdfFlow::filter' by 'dplyr::filter' when
## loading 'flowWorkspace'

parse the workspace into gs

ws <- open_flowjo_xml("~/rglab/workspace/CytoML/wsTestSuite/allstats.wsp")
gs <- flowjo_to_gatingset(ws, name = 1, path = "~/rglab/workspace/CytoML/wsTestSuite/Cytotrol/NHLBI/Tcell/")
gs_get_pop_paths(gs, path = "auto")

## [1] "root"        "Lymphocytes" "CD3"         "cd4"         "HLA"        
## [6] "cd8"

list all the available stats for the entire gating tree

gh <- gs[[1]]
gh_pop_stats_ls(gh)

##  [1] "Count"            "SD"               "Robust CV"        "Mode"            
##  [5] "Median Abs Dev"   "40% Percentile"   "Median"           "Robust SD"       
##  [9] "Mean"             "Geometric Mean"   "Freqoftotal"      "Freqofgranparent"
## [13] "CV"               "Freqof"           "Freqofparent"

list all the available stats for the individual pops

gh_pop_stats_ls(gh, "CD3")

## [1] "Count"

print the stats

default is count

gs_pop_stats_print(gs)

##                            sample         pop  value
## 1: CytoTrol_CytoTrol_1.fcs_119531        root 119531
## 2: CytoTrol_CytoTrol_1.fcs_119531 Lymphocytes  67633
## 3: CytoTrol_CytoTrol_1.fcs_119531         CD3  51450
## 4: CytoTrol_CytoTrol_1.fcs_119531         cd4  31640
## 5: CytoTrol_CytoTrol_1.fcs_119531         HLA  30036
## 6: CytoTrol_CytoTrol_1.fcs_119531         cd8  14934

gs_pop_stats_print(gs, "cd4")

##                            sample pop value
## 1: CytoTrol_CytoTrol_1.fcs_119531 cd4 31640

select different stats, default is cytolib value

gs_pop_stats_print(gh, type = "Freqofparent")

##                            sample pop     value
## 1: CytoTrol_CytoTrol_1.fcs_119531 HLA 0.9493047

the columns are stats specific

gs_pop_stats_print(gh, type = "Freqof")

##                            sample    ancestor pop     value
## 1: CytoTrol_CytoTrol_1.fcs_119531 Lymphocytes HLA 0.4441027

channel-specific stats are not computed by default

gs_pop_stats_print(gs, "HLA", type = "Median")

##                            sample     channel pop value
## 1: CytoTrol_CytoTrol_1.fcs_119531 Comp-B710-A HLA   NaN

gs_pop_stats_print(gs, "HLA", type = "Median", xml = T)

##                            sample     channel pop    value
## 1: CytoTrol_CytoTrol_1.fcs_119531 Comp-B710-A HLA 17778.66

to invoke the stats computing

gs_pop_stats_compute(gs)
gs_pop_stats_print(gs, "HLA", type = "Median")

##                            sample     channel pop    value
## 1: CytoTrol_CytoTrol_1.fcs_119531 Comp-B710-A HLA 17809.85

compare stats

gh_pop_stats_compare(gs, "HLA", type = "Median")

##      type     channel  cytolib      xml node
## 1: Median Comp-B710-A 17809.85 17778.66  HLA

gh_pop_stats_compare(gh)

##     type cytolib    xml        node
## 1: Count  119531 119531        root
## 2: Count   67633  67633 Lymphocytes
## 3: Count   51450  51450         CD3
## 4: Count   31640  31640         cd4
## 5: Count   30036  30036         HLA
## 6: Count   14934  14934         cd8

render it to legacy format for backward compatibility

gh_pop_stats_compare(gh, legacy = T)

##    openCyto.count xml.count        node
## 1:         119531    119531        root
## 2:          67633     67633 Lymphocytes
## 3:          51450     51450         CD3
## 4:          31640     31640         cd4
## 5:          30036     30036         HLA
## 6:          14934     14934         cd8

% is currently concatenated to the type (otherwise we need to append extra column for all the rest stats)

gh_pop_stats_compare(gh, type = "40% Percentile")

##              type     channel  cytolib      xml node
## 1: 40% Percentile Comp-B710-A 16643.83 16612.87  HLA

How to format columns for multiple stats that has different attr (channel vs ancestor)

gh_pop_stats_compare(gh, type = c("Median", "CV"))

## Warning in if (type == "Freqof") col.attr <- "ancestor" else col.attr <-
## "channel": the condition has length > 1 and only the first element will be used

##      type     channel     cytolib        xml node
## 1:     CV Comp-B710-A    26.27054    26.2696  HLA
## 2: Median Comp-B710-A 17809.85156 17778.6641  HLA

Add stats

prevent invalid stats to be added

stat <- list("new_stats")
gs_pop_stats_add(gs, "cd4", stat)

## Error in gs_pop_stats_add(gs, "cd4", stat): stat must be a valid 'cyto_stats' object generated by 'cyto_stats()' constuctor

cyto_stats()

validity checks of constructor

stats <- cyto_stats(type = "Freqof",  channel = c("Comp-V450-A"))

## Error in .cyto_stats(tp, ch, an, pc): channel argument is not applicable for Freqof

stats <- cyto_stats(type = "Median")

## Error in .cyto_stats(tp, ch, an, pc): channels argument is required for Median

construct multi-stats within one call

stats <- cyto_stats(type = c("CV", "Median"),  channel = c("Comp-V450-A", "Comp-G560-A"))
stats

## [[1]]
## cyto_stats 
## type:  CV 
## attr:  Comp-V450-A 
## 
## [[2]]
## cyto_stats 
## type:  CV 
## attr:  Comp-G560-A 
## 
## [[3]]
## cyto_stats 
## type:  Median 
## attr:  Comp-V450-A 
## 
## [[4]]
## cyto_stats 
## type:  Median 
## attr:  Comp-G560-A

add stats

gh_pop_stats_ls(gh, "cd4")

## [1] "Count"

gs_pop_stats_add(gs, "cd4", stats)
gh_pop_stats_ls(gh, "cd4")

## [1] "Median" "CV"     "Count"

gs_pop_stats_compute(gs, "cd4")
gh_pop_stats_compare(gh, "cd4", type = c("CV", "Median"))

## Warning in if (type == "Freqof") col.attr <- "ancestor" else col.attr <-
## "channel": the condition has length > 1 and only the first element will be used

##      type     channel    cytolib xml node
## 1:     CV Comp-V450-A   38.62045 NaN  cd4
## 2: Median Comp-V450-A 6426.73584 NaN  cd4
## 3:     CV Comp-G560-A   91.61989 NaN  cd4
## 4: Median Comp-G560-A 1870.52136 NaN  cd4

remove stats

# gs_pop_stats_remove(gs, "cd4", type = "CV")
#gh_pop_stats_ls(gh, "cd4")