combineMultFilterNAimput: Combine Multiple Filters On NA-Imputed Data

In wrProteo: Proteomics Data Analysis Functions

Combine Multiple Filters On NA-Imputed Data

Description

In most omics data-analysis one needs to employ a certain number of filtering strategies to avoid getting artifacts to the step of statistical testing. This function takes both the origial data and the NA-imputed data to create several different filters (number of NAs, min abundance, ..) and finally combines them. The filter-component aiming to take away the least abundant values (using the imputede data) can be fine-tuned by the argument abundThr.

Usage

combineMultFilterNAimput(
  dat,
  imputed,
  grp,
  useComparison = "all",
  experSetup = NULL,
  annDat = NULL,
  abundThr = NULL,
  colRazNa = NULL,
  colTotNa = NULL,
  minSpeNo = 1,
  minTotNo = 2,
  maxGrpMiss = 1,
  ratMaxNA = 0.8,
  silent = FALSE,
  debug = FALSE,
  callFrom = NULL
)

Arguments

dat	(matrix or data.frame) main data (may contain NA)
imputed	(character) same as 'dat' but with all NA imputed
grp	(character or factor) define groups of replicates (in columns of 'dat')
useComparison	(character or matrix) argument allowing to specify which pairwise comparions sould be performed, default useComparison=NULL will run all pairwise comparisons; may be character as result of combining two group-names (from argument grp) (eg 'A-B', beware, the separator may not appear inside group-names) or 2-column matrix of group-names (thus, without separator; names combined with separator may be shown as rownames) or 2-column matrix of indexes to sorted group-names; default setting of useComparison='all' will select all possible pairwise combinations
experSetup	(list) optional esperimental setup (list with $ind, $pwGrpIndex, $sep and $pwGrpNa as obtained using getPairwiseSetup)
annDat	(matrix or data.frame) annotation data (should match lines of 'dat')
abundThr	(numeric) optional threshold filter for minimumn abundance
colRazNa	(character) if razor peptides are used: column name for razor peptide count
colTotNa	(character) column name for total peptide count
minSpeNo	(integer) minimum number of specific peptides for maintaining proteins
minTotNo	(integer) minimum total ie max razor number of peptides
maxGrpMiss	(numeric) passd to presenceFilt : at least 1 group has not more than this number of NAs (otherwise marke line as bad)
ratMaxNA	(numeric) passd to presenceFilt : at least 1 group below this content of NA values
silent	(logical) suppress messages
debug	(logical) additional messages for debugging
callFrom	(character) allows easier tracking of messages produced

Details

The main filtering step compares the means for each group and line (from argument imputed), at least one group-mean has to be superior the threshold (abundThr, based on hypothesis that if all conditions represent extrememy low measures their differential may not be determined with certainty).

The filter addressing the number of missing values (NA) uses the original data (argument dat), the arguments colTotNa, minSpeNo and minTotNo are used during this step. Basically, this step allows defining a minimum content of 'real' (ie non-NA) values for further considering the measurements as reliable. This part uses internally presenceFilt for filtering elevated content of NA per line.

Finally, this function combines both filters (as matrix of FALSE and TRUE) on NA-imputed and original data and retruns a vector of logical values if corresponding lines passe all filter criteria.

Value

This function returns a list with $data (unmodified input from dat), $nNA, $randParam, $NAneigLst, $seed, $filt (the reommended filtering : TRUE for keeping value), and $annot

See Also

presenceFilt

Examples

set.seed(2013)
datT6 <- matrix(round(rnorm(300)+3,1), ncol=6,
  dimnames=list(paste0("li",1:50), letters[19:24]))
datT6 <- datT6 +matrix(rep(1:nrow(datT6),ncol(datT6)), ncol=ncol(datT6))
datT6[6:7,c(1,3,6)] <- NA
datT6[which(datT6 < 11 & datT6 > 10.5)] <- NA
datT6[which(datT6 < 6 & datT6 > 5)] <- NA
datT6[which(datT6 < 4.6 & datT6 > 4)] <- NA
datT6b <- matrixNAneighbourImpute(datT6, gr=gl(2,3))
datT6c <- combineMultFilterNAimput(datT6, datT6b, grp=gl(2,3), abundThr=2)
head(datT6c$filt)
## with custom choice of comparisons :
gr2 <- gl(2, 3, labels=c("A","B"))
datT6d <- combineMultFilterNAimput(datT6, datT6b, grp=gr2, useCom="B-A", abundThr=2)
head(datT6d$filt)

wrProteo documentation built on July 24, 2026, 1:06 a.m.