Development of a model-based inference approach to detect malfunctioned components in biological systems from clinical data

Detection of malfunctioned reactions or preexisting molecules from the clinical data is essential for drug selection to treat human diseases. Existing approaches based upon Boolean-models or data-driven techniques either ignore the transient information or the detailed description of the reaction networks. In order to address this, a kinetic-model based approach is developed in this work to infer the malfunctioned reactions/preexisting molecules by quantifying the similarity between the output profiles from the malfunctioned model and the profiles shown in the clinical data. The developed approach was tested for four abnormal clinical conditions in IL-6 signaling pathway that include the up/down regulation of single reaction rate constants and up/down regulation of single preexisting molecules. The results show that the developed approach was able to successfully identify the malfunctioned reactions/preexisting molecules from the clinical data. It was found that the developed approach was noise-robust and that it was able to obtained unique solution for the fault in the network from the two measured outputs (i.e., nuclear STAT3 and SOCS3) in the clinical data.