Procedure for developing linear and Bayesian classification models based on immunosensor measurements

Abstract

A protocol for the creation of a set of classification models was developed to differentiate between biological samples based on immunosensor measurements. For this paper, data was gathered using Au Quartz Crystal Microbalance with Dissipation (QCM-D) sensors inoculated with an alkanethiol self-assembling monolayer functionalized for the detection of pAkt, γH2AX, β-Actin, and FITC antigen expression. Oropharyngeal cancer lysate samples, both positive (SCC47) and negative (TU212) for high risk human papillomavirus (HPV16), were used to gather the classification model training data set. Subsequently, linear and Bayesian classifiers were formulated based on the feature values and defined linear discriminant functions. The following study distinguishes between HPV-positive and HPV-negative cell lines, yet these guidelines can be utilized for different immunosensor platforms and disease diagnosis.