Computational protocol: Comparing Imputation Procedures for Affymetrix Gene Expression Datasets Using MAQC Datasets

Similar protocols

Protocol publication

[…] We use the “mas5calls” function detailed in Affymetrix [] from the “affy” package [] to make the detection calls. Using this function, we get a present, marginal, or absent call for each probe set in each array. For every sample, probe sets were filtered based on the present calls where probe sets that were present in all 5 replicates of a given pool and a given site were retained for further analyses. Probe sets that were detected as absent or marginally present in 1 or more replicates of a sample were removed. This creates a complete expression matrix for each site and pool combination. The SimpleAffy R package has methods for quality control metrics on Affymetrix arrays []. One metric is percent present call which calculates the percentage of present probe sets in each array. Using this metric, we calculate the percent present calls for all 120 arrays separately and then average the percentages over the 5 replicates for each sample and each site. […]

Pipeline specifications