I was reading https://www.cruk.cam.ac.uk/wp-content/uploads/2021/02/Quantitative-ChIPseq-Workshop-1.html, which mentions that "More complex schemes, such as forming a consensus peak set separately from the replicates for each condition and then taking the union of these, are also possible."

Is there guidance somewhere to do this? How can this be done with Diffbind?

Thank you,
Ian