Description Usage Arguments Value References See Also Examples
Fast Linear Step Up procedure of Benjamini<e2><80><93>Hochberg FDR method
1 |
pvalues |
A vector of p values. |
alpha |
The desired significance level, the default is 0.05. |
algorithm |
Value of 1 or 2, 1 indicating algorithm 1 in the reference paper for a single chunk without sorting the p values; 2 indicating algorithm 2 in the reference paper for two or more chunks of p values. |
mtg |
Global number of p values in the problem (the sum of all p values in all chunks), should be NA if algorithm is 1. |
mtc |
Number of p values in current chunk (should be less than mtg), should be NA if algorithm is 1. |
The result is a vector of significnat p values at the significance level of alpha.
The "fastLSU" package is created based on Vered Madar and Sandra Batista<e2><80><99>s work. For more details:
Madar V, Batista S. FastLSU: a more practical approach for the Benjamini-Hochberg FDR controlling procedure for huge-scale testing problems. Bioinformatics 2016; 32:1716-23.
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 | set.seed(111)
#simulate p-values for example;
pvals.sim = runif(30000)
Bsig = 1-rbinom(30000,1, .02)
Bsig[Bsig==0] = .0001
pvals.sim = pvals.sim*Bsig
# Example of a single chunk
results.allchunks = fastLSU(pvalues=pvals.sim,alpha=0.1, algorithm = 1)
# Example of 2 chunks of p-values (1st of 10,000, 2nd of 20,000)
pv.chunk1 = pvals.sim[1:10000]
pv.chunk2 = pvals.sim[10001:30000]
# Step 1:
results.chunk1 = fastLSU(pvalues=pv.chunk1,alpha=0.1, algorithm = 2,
mtc=length(pv.chunk1),mtg=length(pvals.sim))
# Step 2:
results.chunk2 = fastLSU(pvalues=pv.chunk2,alpha=0.1, algorithm = 2,
mtc=length(pv.chunk2),mtg=length(pvals.sim))
# Step 3: # get final candidancy
cand.pvals = c(results.chunk1,results.chunk2)
# keep mtc equal to mtg, the fastLSU function already considers length(cand.pvals)
results.final = fastLSU(pvalues=cand.pvals,alpha=0.1, algorithm = 2,
mtc=length(pvals.sim),mtg=length(pvals.sim))
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.