Entering edit mode
Anna Christine Reisetter
▴
20
@anna-christine-reisetter-6692
Last seen 9.6 years ago
Hi,
I'm using the ComBat package within the MSPrep R package to normalize
some metabolomic data. I receive the following error when trying to
run MSPrep:
Error in while (change > conv) { : missing value where TRUE/FALSE
needed
I traced this error to the ComBat function and within it, the it.sol
function, the code for which I obtained from the ComBat website
(http://www.bu.edu/jlab/wp-assets/ComBat/Download.html) and have
pasted below.
it.sol <- function(sdat,g.hat,d.hat,g.bar,t2,a,b,conv=.0001){
n <- apply(!is.na(sdat),1,sum)
g.old <- g.hat
d.old <- d.hat
change <- 1
count <- 0
while(change>conv){
g.new <- postmean(g.hat,g.bar,n,d.old,t2)
sum2 <- apply((sdat-g.new%*%t(rep(1,ncol(sdat))))^2, 1,
sum,na.rm=T)
d.new <- postvar(sum2,n,a,b)
change <-
max(abs(g.new-g.old)/g.old,abs(d.new-d.old)/d.old)
g.old <- g.new
d.old <- d.new
count <- count+1
}
#cat("This batch took", count, "iterations until
convergence\n")
adjust <- rbind(g.new, d.new)
rownames(adjust) <- c("g.star","d.star")
adjust
}
The MSPrep package passes the following arguments through the
function:
for (i in 1:n.batch) {
temp <- it.sol(s.data[, batches[[i]]], gamma.hat[i, ],
delta.hat[i,
], gamma.bar[i], t2[i], a.prior[i], b.prior[i])
gamma.star <- rbind(gamma.star, temp[1, ])
delta.star <- rbind(delta.star, temp[2, ])
}
Where n.batch=50, the dimensions of s.data are (392, 300), the
dimensions of gamma.hat (g.hat) are (50, 392), the dimensions of
delta.hat (d.hat) are (50, 392), gamma.bar (g.bar) is NULL, a.prior
(a) is NULL, and b.prior (b) is null).
I believe the problem is coming from the fact that gamma.bar is null,
which results because gamma.hat has NaN values.
Could you please advise?
Thank you,
Anna
Anna Reisetter, MS
Statistical Analyst/Programmer
Department of Preventive Medicine
Northwestern University Medical School
680 N. Lake Shore Drive Suite 1400
Chicago, IL 60611
312 503 7039
anna.reisetter@northwestern.edu<mailto:anna.reisetter@northwestern.edu>
[[alternative HTML version deleted]]