Abstract: Provided are methods of characterizing and/or predicting risk associated with a biological sample using thermal stability profiles. The methods include obtaining a thermal stability profile of the sample, using a sensor which detects heat capacity values, applying a classification algorithm to the thermal stability profile, and comparing the results thermal stability data in the database to characterize and/or predict risk of the condition. The methods may additionally include a classification algorithm selected from one or more of logistic regression, support vector machines, Fisher's linear discriminant analysis, modified version of Fisher's linear discriminant analysis (MLDA), and partial least squares.