Fig. 1
From: Redundancy analysis allows improved detection of methylation changes in large genomic regions
Precision and power of RDA, DMRcate, and blockFinder for simulated sets of 40 samples. To be considered a true positive, DMRcate regions should comprise at least 50% of the simulated region. A DMRcate region including CpGs outside the DMR was considered as a false positive. DMRcate parameters were set to preserve the default smoothing window while allowing for DMRs as big as our target DMRs. Each sub figure represents a different scenario and each shape a different size of the simulated DMR. Results were derived from 200 simulations