-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmyDiffAbuFunctions.R
548 lines (423 loc) · 23.9 KB
/
myDiffAbuFunctions.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
#' -----------------------------------------------------------------------------------------------------------------------------
# --------------------------------------- my DiffAbu plotter function --------------------------------------
myDAplotter <- function(DA_obj, fcts, rand=NULL, corr=NULL, transf="LOG", effZ=1, sig = 0.2, ntax=25, check=F,
fc.color=NULL, fc.ncol=NULL, fc.nrow=NULL, colVector = myColors$discrete){
DA_resMTX <- DA_obj$DA_resMTX
amp_obj <- DA_obj$amp_obj
agg_at <- DA_obj$agg_at
DA_type <- DA_obj$DA_type
##' flt &fmt DA_resMTX
if(DA_type=="MaAsLin2"){coef_col = "coef"; sig_col="qval"; metCols <- NULL } # "metadata"
if(DA_type=="ALDEx2" ){coef_col = "effect"; sig_col="p.adj"; metCols <- NULL }
if(DA_type=="DESeq2" ){coef_col = "log2FoldChange"; sig_col="padj"; metCols <- NULL }
### lets rename columns to keep them consistent
colnames(DA_resMTX)[match(coef_col,colnames(DA_resMTX) ) ] <- "EffectSize"
colnames(DA_resMTX)[match(sig_col, colnames(DA_resMTX))] <- "Adj.Significance"
# colnames(DA_resMTX)[match("metadata", colnames(DA_resMTX)) ] <- fcts[[1]]
#### select significance thresholds
taxCols <- c("TAX", "Taxon", "Class", "Phylm")
DA.Cols <- c("EffectSize", "Adj.Significance") #EffectSize, qValue
DA_resMTX.sig <- DA_resMTX %>% filter(Adj.Significance < sig & abs(EffectSize) > effZ ) %>% arrange(desc(abs(EffectSize))) %>%
relocate(TAX, .before=Taxon) %>% dplyr::select(all_of(c(DA.Cols, taxCols))) #metCols,
DA.sigTAX=row.names(DA_resMTX.sig) ##%>% sort()
DA_resMTX.sig_kbl <- unique(DA_resMTX.sig) %>% select(-c(TAX)) %>%
mutate(across(EffectSize, \(x) round(x, 2) ) ) %>% #mutate(across(EffectSize, round, 2) ) %>%
mutate(Adj.Significance = sprintf("%.2e", Adj.Significance)) %>% kable("html", digits=150) %>% kable_classic_2(full_width=F) %>%
kable_styling("striped") %>% scroll_box(width = "70%", height = "500px")
info = c( DA_type ,"total differential TAXa ~", fcts[[1]],"(at ", agg_at, "level) : ", dim(DA_resMTX.sig)[1], ".", '\n',
"Significance threshold < ", sig, '\n',
"Effect size threshold >", effZ, '\n')
### getting the flt & fmt DA res out (just in case here)
res_obj <- list(DA.sigTAX, DA_resMTX.sig, DA_resMTX.sig_kbl)
objNames <- c("DA.sigTAXa", "DA.sig.tab", "DA.sig_kbl")
res_inf <- list(info, sig, effZ, ntax, DA_type, agg_at)
infNames <- c("info", "sig", "effZ", "ntax", "DA_type", "agg_at")
all <- do.call(c, list(res_obj, res_inf))
names(all)<- c(objNames, infNames)
##' plotting volcanos
vlc.title = paste("TAXa changes ~", fcts[[1]])
vlc.subtitle = DA_type
vlc.caption = paste ("total = ", nrow(DA_resMTX), agg_at)
library(EnhancedVolcano)
volcano.plot <- EnhancedVolcano(DA_resMTX, x="EffectSize", y="Adj.Significance", lab=DA_resMTX$Taxon,
FCcutoff= effZ, pCutoff=sig, legendPosition = "bottom",
caption = vlc.caption, captionLabSize = 10,
ylab="-Log10(Adj.Significance)", titleLabSize = 12, title=vlc.title, xlab="EffectSize",
subtitle = vlc.subtitle, legendLabSize=10 , axisLabSize = 10, labSize = 3);
# volcano.plot
##' -------- flt & fmt amp_obj
##' agglomerated amp used from DA_obj
##' agglomerate if needed
# if(agg_at!= "OTU"){
# print(paste("aggregating TAXa to", agg_at, "level"))
# source(paste0("~/Documents/MyApps/RScripts/myRscripts/dataClean.R"))
# amp_obj <- amp_agg_abund(amp_obj, agg_at = agg_at)
# # amp_obj
# # print(amp_obj$tax %>% head(12))
# }
counts <- amp_obj$abund
tax <- amp_obj$tax
meta <- amp_obj$metadata
meta_sub <- meta %>% select(any_of(c(fcts, rand, corr))) %>% mutate_all(factor)
if (transf=="LOG"){
# the log2(abd) plot: log-ratios are captured
txfCounts <- log((counts+1)/(rowSums(counts)+dim(counts)[2]))
txfCounts %<>% t() %>% as.data.frame() %>% select(DA_resMTX.sig$TAX)
}else{
txfCounts <- (counts+1)/(rowSums(counts)+dim(counts)[2])
txfCounts %<>% t() %>% as.data.frame() %>% select(DA_resMTX.sig.sig$TAX)
}
##' try DF4plot
if(check==T){
print("checkpoint DF4plot")
print(DA_resMTX.sig %>% head)
}
topSigTAX <- DA.sigTAX[1:ntax] %>% na.omit() %>% c()
df4plot0<-NULL
try({
for(i in topSigTAX){
tmp<-NULL
# meta_sub <- meta %>% select(any_of(c(fcts, rand, corr))) # is same as meta
tmp<-data.frame(txfCounts[,i], meta_sub, rep(i),
rep(paste0(DA_resMTX.sig$Taxon[DA_resMTX.sig$TAX == i], " (", i,")" ) ),
# rep(paste0("EffectSize = ", sprintf("%.4g", top_sig_tax$effect[top_sig_tax$TAX== i]),
# "; qval.BH = ", sprintf("%.1e", top_sig_tax$glm.eBH[top_sig_tax$TAX==i]) ) ),
rep( DA_resMTX.sig$EffectSize[DA_resMTX.sig$TAX== i]) ,
rep( DA_resMTX.sig$Adj.Significance[DA_resMTX.sig$TAX==i]) )
if(is.null(df4plot0)){df4plot0<-tmp} else { df4plot0<-rbind(df4plot0,tmp)}
}
# print(tmp)
#the value is log(relAbu) of TAXa in each Sample
colnames(df4plot0)<-c("Value", c(fcts, rand, corr), "TAX" , "Taxon", "EffectSize", "Adj.Significance");
df4plot0$Taxon %<>% factor( levels=unique(df4plot0$Taxon) )
})
##' plotting FC/EffZ plots
# df4plot0$Taxon %>% unique()
df4plot <- df4plot0 %>% mutate(SigVal=paste0("Eff.size=", formatC(EffectSize, digits=2),
"; pVal =", formatC(Adj.Significance, format="E", digits=2)) )
# fc.color=NULL; fc.ncol=4; fc.nrow=NULL;
if(is.null(fc.color)){fc.color = fcts[[1]] }
fc.title = paste(DA_type, "top", ntax, "differential TAXa ( at",agg_at, "level) b/w", fcts[[1]], "groups")
if (transf=="LOG"){ ylabel = "Log2(Relative abundance)" }else{ylabel = "Relative abundance" }
## info about the aes_string() replacement: https://www.tidyverse.org/blog/2018/07/ggplot2-tidy-evaluation/
foldchange.plot <- ggplot(unique(df4plot), aes_string(fcts[[1]], "Value", color= fc.color )) + # #aes(vars(fcts[[1]]), "Value", color= fc.color )
ylab(ylabel) + labs(title=fc.title, color=fc.color) +
geom_boxplot(outlier.size = 0) + xlab(fcts[[1]]) +
geom_jitter(position = position_jitterdodge(), alpha=0.3) +
facet_wrap( ~ Taxon + SigVal, scales="free_x", #labeller = label_wrap_gen(width=45),
ncol=fc.ncol, nrow=fc.nrow ) + theme_bw() + scale_color_manual(values=colVector) +
theme( axis.text.x=element_text(angle=0,hjust=0.5,vjust=0.5, size = 14), legend.position = "top",
strip.text.x = element_text(size = 12, colour = "black")) #, angle = 90
foldchange.plot
res_plt<- list(volcano.plot, foldchange.plot, df4plot)
pltNames <- c("VolcanoPlot", "EffSizePlot", "DF4plot")
all<- do.call(c, list(res_plt, res_obj, res_inf))
names(all) <- c( pltNames, objNames, infNames)
return(all)
}
#' -----------------------------------------------------------------------------------------------------------------------------
# ----------------------------- my DESeq2 function --------------------------------------
##' e.g. use: DA_objDSQ <- myDESeq2(ampHPV, fcts = fcts, agg_at = "OTU")
##' For the correction design: **add the biological factors to the fml(fcts),**
##' ** but the stat confounders (demographics, batches, etc) go in the correction factor**
myDESeq2<-function(amp_obj, fcts, corr=NULL, agg_at= "OTU", check=F, reffLvL=NULL){
##' agglomerate if needed
if(agg_at!= "OTU"){
print(paste("aggregating TAXa to", agg_at, "level"))
source(paste0("~/Documents/MyApps/RScripts/myRscripts/dataClean.R"))
amp_obj <- amp_agg_abund(amp_obj, agg_at = agg_at)
# amp_obj
# print(amp_obj$tax %>% head(12))
}
counts <- amp_obj$abund
tax <- amp_obj$tax
meta <- amp_obj$metadata
meta_sub <- meta %>% select(any_of(c(fcts, corr))) %>% mutate_all(factor)
# if(is.null(ref_lvl)){reff=NULL}else{ reff <- paste0( ref_lvl , ",", fcts[[1]] ) }
# reff
#'------------------------ data PREP ---------------------------------------------
countdata = round(as(counts, "matrix"), digits = 0) +1
DA_type="DESeq2"
#'------------------------ DDS calculation ------------------------
if(is.null(corr)){
# fml <- as.formula(paste(" ~", fcts[[1]]) ) ## paste(fcts, collapse = "+") ) # !!!!!!!!!!!!!!!!!!!
fml <- as.formula(paste(" ~", paste(rev(fcts), collapse="+")))
redfml <- as.formula(paste("~1") )
}else{
fml <- as.formula(paste(" ~ ", paste(rev(corr), collapse = " + "), " + ", paste(rev(fcts), collapse="+") ) ) # paste(rev(fcts), collapse = " + "))
redfml <- as.formula(paste("~1 + ", paste(rev(corr), collapse = "+") ) )
}
library(DESeq2)
library(BiocParallel)
threads=detectCores()-4
register(MulticoreParam(threads))
dds <- DESeqDataSetFromMatrix(countdata, meta_sub, fml)
dsqRAW = DESeq(dds, test="LRT", reduced = redfml, parallel = T);
# dispPlot <- DESeq2::plotDispEsts(dsqRAW)
#' -------------------------- extracting DDS results ---------------------------------------------
#' If results is run without specifying contrast or name (name is for continuous variables), \
#' it will return the comparison of the last level of the last variable in the design formula \
#' over the first level of this variable. (log2(lastGroup/firstGroup) =>
#' **firstGroup listed in the levels of a factor, will be considered RefflvL (the Control/Negative/untreated))**
#' To set contrast manually, set contrast=c("condition","treated","untreated.Control.Negative")
dsqRES = results(dsqRAW, cooksCutoff = F)#, contrast = reffLvL)
dsqRES <- dsqRES[order(dsqRES$padj),];
# if(check==T){ print(head(res, 4)) }
#' ----------------------- adding TAXnames to dsqRES ---------------------------------------------
if(agg_at=="OTU"){lRank="Species"}else{lRank=agg_at}
dsqRES_tax = cbind(as.data.frame(dsqRES) , # as.matrix(countdata[rownames(dsqRES), ]),
Taxon = tax[[lRank]][match(rownames(dsqRES), rownames(tax) ) ],
Class = tax$Class[match(rownames(dsqRES), rownames(tax) ) ] ,
Phylm = tax$Phylum[match(rownames(dsqRES), rownames(tax) )] )
if(check==T){
print(fml); print(redfml)
print(head(dsqRES, 4))
y<- ncol(dsqRES_tax); y1=y-5
print(dsqRES_tax[1:5, y1:y] )
}
##' --- output formatting
dsqRES_tax <- dsqRES_tax %>% arrange(desc(abs(log2FoldChange))) %>%
mutate(TAX=rownames(.)) %>% relocate(TAX, .before="Taxon")
all<- list(dsqRES_tax, agg_at, DA_type, amp_obj, fcts, corr, dsqRAW, dsqRES)
names(all) <- c("DA_resMTX", "agg_at", "DA_type", "amp_obj", "fcts", "corr", "dsqRAW", "dsqRES")
return(all)
}
#' -----------------------------------------------------------------------------------------------------------------------------
# -------------------------------------- wrapper Maaslin function -------
##' e.g. use DA_objMSN<- wrpMaAsLin2(ampHPV, fcts, maasDIR=DIR, model="LM", agg_at = "Genus", rand = rand)
wrpMaAsLin2 <- function(amp_obj, fcts, rand = NULL, maasDIR= NULL, min_prevalence=0.1, min_abundance=0.0,
reff_lvl=NULL, agg_at = "OTU", model = "all"){
##' agglomerate if needed
if(agg_at!= "OTU"){
print(paste("aggregating TAXa to", agg_at, "level"))
source(paste0("~/Documents/MyApps/RScripts/myRscripts/dataClean.R"))
amp_obj <- amp_agg_abund(amp_obj, agg_at = agg_at)
# amp_obj
# print(amp_obj$tax %>% head(12))
}
counts <- amp_obj$abund
tax <- amp_obj$tax
meta <- amp_obj$metadata
meta %<>% select(any_of(c(fcts, rand))) %>% mutate_all( factor)
library(Maaslin2)
DA_type <- "MaAsLin2"
if(model %in% c("LM", "CPLM","NEGBIN","ZINB") ){norm=NULL; transf=NULL } # wrapper will auto do all combinations
if(model=="all"){ model=NULL; norm=NULL; transf=NULL } # wrapper will auto do all combinations
#' ------------------ the wrapper code --------------------
# print(list(model, norm, transf) )
source(paste0(myRpath, "otherFunctions.R"))
phy <- amp2phy(amp_obj)
source(paste0(myRpath, "maaslin_wrapper_AA.R"))
start.time <- Sys.time()
suppressMessages( suppressWarnings( {
maasOUT = multimodel_maaslin(phy=phy, fixed_effects=fcts, output_dir=maasDIR, random_effects= rand,
min_abundance = min_abundance , min_prevalence = min_prevalence, #needs % at >2 samples
standardize=F, model=model, txf = transf, norm = norm)
}) )
end.time <- Sys.time()
time.taken <- round(end.time - start.time); print(time.taken)
# print(maasOUT %>% head() )
##' --- output formatting
##' feature to TAX & rownames
maasOUT <- maasOUT %>% filter(metadata %in% fcts[[1]]) %>% select(-all_of(c("metadata", "value", "name") ) ) %>%
mutate(TAX=feature) %>% relocate(TAX, .after=last_col() )
##' in cases fct[main] has 2+ categories, we get pair-wise comparisons to REFF_level => duplicated features =>
##' => problem with assigning row names =>
##' will extract only distinct(feature) with higher abs(coef)
maasOUT <- maasOUT %>% arrange(desc(abs(coef)) ) %>% distinct(feature, .keep_all = TRUE) %>%
column_to_rownames("feature") %>% arrange(qval)
if(agg_at=="OTU"){lRank="Species"}else{lRank=agg_at}
# maasOUTt = cbind(maasOUT,
# Taxon = tax[[lRank]][match(rownames(maasOUT), rownames(tax) ) ],
# Class = tax$Class[match(rownames(maasOUT), rownames(tax) ) ],
# Phylm = tax$Phylum[match(rownames(maasOUT), rownames(tax) ) ] )
maasOUTt = cbind(maasOUT,
Taxon = tax[[lRank]][match(maasOUT$TAX, rownames(tax) ) ],
Class = tax$Class[match(maasOUT$TAX, rownames(tax) ) ],
Phylm = tax$Phylum[match(maasOUT$TAX, rownames(tax) ) ] )
all <- list(maasOUTt, agg_at, DA_type, amp_obj, fcts, rand)
names(all) <- c("DA_resMTX", "agg_at", "DA_type", "amp_obj", "fcts", "rand")
return(all)
}
#' ---------------------------------------------------------------------------------------------------------------------------
# --------------- my Maaslin funciton ----------------------------------------------------------
myMaAsLin2 <- function(amp_obj, fcts, rand = NULL, maasDIR= "maasOUT", minPrev=0.1, minAbu=0.0,
reffLvL=NULL, agg_at = "OTU", model = "LM", transf="LOG", norm="CLR"){
##' agglomerate if needed
if(agg_at!= "OTU"){
print(paste("aggregating TAXa to", agg_at, "level"))
source(paste0("~/Documents/MyApps/RScripts/myRscripts/dataClean.R"))
amp_obj <- amp_agg_abund(amp_obj, agg_at = agg_at)
# amp_obj
# print(amp_obj$tax %>% head(12))
}
counts <- amp_obj$abund
tax <- amp_obj$tax
meta <- amp_obj$metadata
meta %<>% select(any_of(c(fcts, rand))) %>% mutate_all( factor)
library(Maaslin2)
DA_type <- "MaAsLin2"
####### ---- direct maaslin run ---
##' To use reffLvL do: revLvL = paste0('FCT[[1]], REFlvl'). e.g. **reffLvL=paste0("HPVstatus,HPVneg")**
# met$Groups=paste(met[[fxd_effects[[1]]]], met[[fixed_effects[[2]]]], sep="-")
# if(!is.null(reff_lvl)){ reff_lvl <- c(paste0(fcts[[1]], ",", reff_lvl ) ) }
# threads=detectCores()-4
# register(MulticoreParam(threads))
print("using 1 model maaslin ")
start.time <- Sys.time()
# suppressMessages( suppressWarnings(
maasOUTa = Maaslin2::Maaslin2(input_data = counts, input_metadata = meta, output = maasDIR,
fixed_effects = fcts,
random_effects = rand,
reference= reffLvL,
min_abundance = minAbu,
min_prevalence = minPrev, #neeeds at least 2 samples, so 10% is safest
normalization= norm, transform= transf, analysis_method = model,
plot_heatmap=F, plot_scatter=F, save_models = T, cores=1) ###keep cores at 1, fastest!!
# ))
end.time <- Sys.time()
time.taken <- round(end.time - start.time); print(time.taken)
#' Add Poorani's AIC function (June 12)
AIC_forcplm <- function(fit) {
#' extracts aic value from slot of fit
#' is of class cpglm or uses AIC() otherwise
#' @param fit model fit
#' @return AIC value
if(inherits(fit, 'cpglm')) {
return(fit$aic)
} else{
return(AIC(fit))
}
}
AICs <- data.frame(AIC=unlist(lapply(maasOUTa[5]$fits, function(x) tryCatch(AIC_forcplm(x), error=function(e) NA))))
maasOUT <- merge(maasOUTa$results, round(AICs, 2), by.x="feature", by.y=0)
##' --- output formatting
##' feature to TAX & rownames
maasOUT <- maasOUT %>% filter(metadata %in% fcts[[1]]) %>% select(-all_of(c("metadata", "value", "name") ) ) %>%
mutate(TAX=feature) %>% relocate(TAX, .after=last_col() )
##' in cases fct[main] has 2+ categories, we get pair-wise comparisons to REFF_level => duplicated features =>
##' => problem with assigning row names =>
##' will extract only distinct(feature) with higher abs(coef)
maasOUT <- maasOUT %>% arrange(desc(abs(coef)) ) %>% distinct(feature, .keep_all = TRUE) %>%
column_to_rownames("feature") %>% arrange(qval)
if(agg_at=="OTU"){lRank="Species"}else{lRank=agg_at}
maasOUTt = cbind(maasOUT,
Taxon = tax[[lRank]][match(rownames(maasOUT), rownames(tax) ) ],
Class = tax$Class[match(rownames(maasOUT), rownames(tax) ) ],
Phylm = tax$Phylum[match(rownames(maasOUT), rownames(tax) ) ] )
all <- list(maasOUTt, agg_at, DA_type, amp_obj, fcts, rand)
names(all) <- c("DA_resMTX", "agg_at", "DA_type", "amp_obj", "fcts", "rand")
return(all)
}
#' -----------------------------------------------------------------------------------------------------------------------------
# ---------------------------------- ALDEX function (fancy) --------------------------------------
##' can do 1 factor with 2 levels in KW or GLM mode (GLM mode depending on sufficient contrast)
##' can do 1 factor with 2+ levels in GLM mode
##' can do 1 factor with >=2 levels with adjustments (corr)
##' is slow even in multi-thread mode, but KW is fastest
##' can agglomerate to TAX level chosen
##' fct1 should be pre-set to list CTRL level first in list of factors
##' works with myDAplotter
myALDEx2 <- function(amp_obj, fct1, corr= NULL, agg_at = "OTU", mode="glm", check=F, denom="all", threads=NULL){
#ensure fct1 & corr levels are listed with CTRLlvl listed first
##' agglomerate if needed
if(agg_at!= "OTU"){
print(paste("aggregating TAXa to", agg_at, "level"))
source(paste0("~/Documents/MyApps/RScripts/myRscripts/dataClean.R"))
amp_obj <- amp_agg_abund(amp_obj, agg_at = agg_at)
# amp_obj
# print(amp_obj$tax %>% head(12))
}
meta <- amp_obj$metadata # %>% mutate_at(., c(fct1, corr), factor) #factorl lvls should be pre-made, CTRL lvl list 1st
counts <- amp_obj$abund
tax <- amp_obj$tax
###' ------------ settings
###' for KW mode ## only for 1 factor with 2 levels (quicker)
if(mode == "kw") { #only for 1 factor with 2 levels
fml <- as.formula(paste0("~", fct1))
if(denom!="all"){denom="iqlr"}
mm <- meta[[fct1]]
if(check==T){print(levels(mm))}
}
###' ------------ settings
###' for GLM mode ## for 1 main factor with multiple levels & corrections
if(mode== "glm"){
if(is.null(corr) ){ # main factor with 2+, but not corrections
fml <- as.formula(paste("~ ", fct1) )
}else{ # main factor with 2+ levels & correction
fml <- as.formula(paste("~ ", fct1, " + ", paste(corr, collapse="+") ))
}
denom = "all" ## iqrl not applicable in GLM mode
refLvL= "RefLvL"
mm <- model.matrix(fml, data=meta) #assumes refLvL listed first
}
DA_type="ALDEx2"
library(ALDEx2)
library(BiocParallel)
if(is.null(threads)) {threads=detectCores()-4 }
register(MulticoreParam(threads))
xx <- aldex.clr(counts, # the rows are SampleNames and Columns are TAXa
mm, # the model matrix
mc.samples=128, # monte carlo (random) samples to use for estimating distribution: 16 or [def: 128]
denom=denom, # only all is accepted for glm
verbose=T, useMC = T)
# row.names(xx@analysisData$DNA103)
###' ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
###' for KW mode #only for 1 factor with 2 levels
if(mode == "kw"){
x.kw <- aldex.kw(xx, verbose = T)
x.eff <- aldex.effect(xx, verbose = T)
x.all <- data.frame(x.kw, x.eff)
x.all %<>% arrange(desc(abs(effect))) %>% dplyr::select(any_of(c("effect", "glm.eBH"))) %>%
mutate(TAX=rownames(.)) %>% dplyr::rename(p.adj = "glm.eBH")
if(check==T){ print(x.all %>% head(20) ) }
}
###' ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
###' for GLM mode
if(mode == "glm"){
print("starting aldex GLM")
# Buggy function aldex.glm().
# Common error: "missing value where TRUE/FALSE needed" @ "verbose[1] == TRUE"?
# Has smth to do with no insufficient contrasts bw groups? Try lowering agg_at level or seqDepth
xx.glm <-aldex.glm(xx, mm) #do not input verbose (errors). Cannot useMC.
# colnames(xx.glm)
xx.glm1 <- xx.glm %>% dplyr::select(contains(fct1) ) %>% ## as w Maaslin we select only the res of main factor
dplyr::select(contains("pval.holm")) %>% ## only pval is impotant. No effect GLM
select(-contains("Intercept"))%>% rename_all(~str_replace(., fct1, paste0(refLvL, "~") ) )
if(check==T){ colnames(xx.glm1) %>% print() }
print("starting aldex effect")
xx.eff <- aldex.glm.effect(xx, verbose = TRUE, useMC = TRUE);
# xx.eff$Group1HPVcarc #%>% conames()
xx.eff1 = data.frame(xx.eff) %>% dplyr::select(contains("effect")) %>% ## only effect is collected here
dplyr::select(contains(fct1)) %>% ## as w Maaslin we select only the res of main factor
rename_all(~str_replace(.,fct1, paste0(refLvL, "~")));
if(check==T){ colnames(xx.eff1) %>% print() }
xx.all = data.frame(xx.glm1, xx.eff1) # we have already pre-select only the res of main factor
if(check==T){ xx.all %>% head(5)%>% print() }
#creating new column with the value.which = max(abs(effect)) & another column with value.which = min(qval)
xx.all1 <- xx.all %>% mutate(effect = apply(dplyr::select(., contains("effect")), 1, \(x) x[which.max(abs(x))] )) %>%
mutate(p.adj = apply(dplyr::select(., contains(".holm")), 1, min)) %>%
select(c("effect", "p.adj"))
x.all <- xx.all1 %>% arrange(desc(abs(effect))) %>% mutate(TAX=rownames(.))
if(check==T){ xx.all1 %>% head(5) %>% print() }
}
if(agg_at=="OTU"){lRank="Species"}else{lRank=agg_at}
x.all = cbind(x.all,
Taxon = tax[[lRank]][match(rownames(x.all), rownames(tax) ) ],
Class = tax$Class[match(rownames(x.all), rownames(tax) ) ],
Phylm = tax$Phylum[match(rownames(x.all), rownames(tax) ) ] )
all <- list(x.all, agg_at, DA_type, amp_obj, fct1, corr)
names(all) <- c("DA_resMTX", "agg_at", "DA_type", "amp_obj", "fcts", "corr")
return(all)
}
#'------------
#' ---------------------------------------------------------------------------------------------------
# ---------------------------------- Compare lists function ----------------------
##' to compare output of sigABU lists
list_compare <-function(list1, list2, taxa_list, agg_at="Species"){
comTax<-try({Reduce(intersect, list(list1, list2) ) })
comNam<- taxa_list[[agg_at]][row.names(taxa_list) %in% comTax]
common <- list(comTax, comNam)
names(common) <- c("comTax", "comNam")
return(common)
}