I’m working with Saccharomyces cerevisiae (budding yeast), a gene-dense organism with a compact 12-Mb genome (approximately 2.4 orders of magnitude smaller than human, 2.3 smaller than mouse), and using MACS3 for ChIP-seq peak calling.
The default values for the --slocal and --llocal parameters in MACS3 are set to 1 kb and 10 kb, respectively. These parameters define the window sizes for the lambda (Poisson) background model used to estimate read coverage enrichment in relation to background noise.
My working assumption is that these defaults are optimized for larger-genome organisms (e.g., human or mouse; please correct me if this is wrong), so I’m wondering if other users adjust these parameters (or others) when working with small-genome organisms like yeast to improve peak calling sensitivity or specificity? Any insights or feedback would be appreciated.
