Skip to contents

Frozen robust-regression hemolysis correction introduced in the single-sample scoring bank (Module B). Performs per-feature robust regression of log-expression on a hemolysis proxy score (and optionally a platelet score), fitting only on training controls. At deployment, the predicted nuisance contribution is subtracted from each feature.

Note: the existing package file hemolysis-correction.R provides fit_hemolysis_prefilter() (a different approach based on marker-ratio gating). This file provides the Module-B fit-and-freeze approach and is named singlesample-hemolysis.R to avoid collision.

Methods provided:

References

Huber PJ. (1964) Robust Estimation of a Location Parameter. The Annals of Mathematical Statistics 35(1): 73–101.