Adapting censored regression methods to adjust for the limit of detection in the calibration of diagnostic rules for clinical mass spectrometry proteomic data
Clinical mass spectrometry-based proteomicslimit of detectioncensored regressionborrowing of informationpredictionvariable selection
