aylafr
December 14, 2022, 11:39am
1
hello,
I was wondering how the order of operations is in maaslin.
If I do a min_prevalence = 0.1 and normalization = “CLR” is one done before the other as default or will it change depending on the order inputted in my code?
Thank you
As you can see in the source, the prevalence is used first (lines 843 then 853-6) before the normalization (lines 884-5).
unfiltered_data <- data
unfiltered_metadata <- metadata
# require at least total samples * min prevalence values
# for each feature to be greater than min abundance
logging::loginfo(
"Filter data based on min abundance and min prevalence")
total_samples <- nrow(unfiltered_data)
logging::loginfo("Total samples in data: %d", total_samples)
min_samples <- total_samples * min_prevalence
logging::loginfo(
paste("Min samples required with min abundance",
"for a feature not to be filtered: %f"),
min_samples
)
# Filter by abundance using zero as value for NAs
data_zeros <- unfiltered_data
data_zeros[is.na(data_zeros)] <- 0
filtered_data <-