Endpoints for slope calculation?

[amwa]

Hi! I'm just starting to use this tool and was wondering - how are the X and Y coordinates of the two points used for slope calculation determined? I tried running the program twice, first on a given set of samples, and next on a smaller subset of the first set, and got noticeably different X/Y coordinates and scaling factors for the same samples, which was confusing to me.

[hongjianjin]

ChIPseqSpikeInFree calculates cutoffs (minFirstTurn, maxLastTurn) based on a group of samples using same ANTIBODY and then uses same cutoffs for the same set of samples to exclude potential noise regions . So exclusion of one sample could change cutoffs , subsequentially affect scaling factor for other samples in the same group. This design is not ideal but reasonable for group-based comparisons for each antibody (at least 3 replicates per group , comparing one vs another group) . It is not meaningful to compare samples in same group or same samples across different runs, or sample/groups from different antibodies.