ChIP-seq normalization, Signal-to-Noise different between samples

Question

[Kelsey Tyssowski]

I have ChIP-seq data for several time points after a stimulus, but it seems like my samples have differences in signal-to-noise ratio. i.e. some samples have a more background reads relative to reads in peaks than others, likely due to technical variation. Is there a standard way to normalize for this?

Maybe something like MAnorm? But does that work when comparing more than two samples?

[Devon Ryan]

Have a look at the SES normalization method in bamCompare from deepTools. It will take SNR into account when normalizing to input.