Ramp Fitting
Description
The step FitRampStep determines the effective slope for a sequence of UTR reads. This is currently implemented with numba.
Algorithm
Three methods are available for fitting the ramp:
1. Correlated Double Sampling (CDS) - The difference between the first (or probably second for most exposures) and last read are used to determine the slope.
2. Multiple Correlated Double Sampling (MCDS) - Similar to CDS, but multiple reads at the begining and end of the ramp are used to determine the slope.
3. Ordinary least squares (OLS) - The entire ramp is fit in a least squares framework. The current algorithm assumes reads are not correlated. Future algorithms will use a generalized least squares framework accounting for the covariance of each read.
Arguments
- input
str|RampModel The input file or ramp model to process.
- method
str Which method to use for ramp fitting. Options are ‘CDS’, ‘MCDS’, or ‘OLS’. Default is ‘OLS’.
- num_coadds
int Number of coadds to use for ramp fitting in MCDS mode. Default is 1 (equivalent to
method='CDS').