id summary reporter owner description type status priority milestone component version resolution keywords cc
425 Implement qQuantile normalizer Jari Häkkinen Jari Häkkinen "The idea is to normalize each column in a matrix to match a target columm (or maybe this should be a target vector).
Matching is done like this for each column:
1. Partition into quantiles, q_i=(i-0,5)/N, i=1, 2, ..., N
1. Calculate the arithmetic mean for each quantile
1. Do the same for the target
1. Calculate the difference between of target and column means. Now we have N differences d_i.
1. Create a cubic spline fit to this difference vector d. The resulting curve is used to recalculate all column values.
a. For values in range [q_i, q_N] we use the cubic spline fit.
a. For values outside the range, q_N a linear extrapolation is used
Linear interpolation simply means a translation here (at least for now).
This method is inspired by the work of Workman ''et al.'' ""A new non-linear normalization method for reducing variability in DNA microarray experiments"", Genome Biol. 2002; 3(9):
PMID: 12225587 [http://genomebiology.com/2002/3/9/research/0048 article at Genome Biology]" enhancement closed major yat 0.5 normalizer trunk fixed