Latent inter-organ mechanism of idiopathic pulmonary fibrosis unveiled by a generative computational approach.

Idiopathic pulmonary fibrosis (IPF) is a chronic and progressive disease characterized by complex lung pathogenesis affecting approximately three million people worldwide. While the molecular and cellular details of the IPF mechanism is emerging, our current understanding is centered around the lung itself. On the other hand, many human diseases are the products of complex multi-organ interactions. Hence, we postulate that a dysfunctional crosstalk of the lung with other organs plays a causative role in the onset, progression and/or complications of IPF. In this study, we employed a generative computational approach to identify such inter-organ mechanism of IPF. This approach found unexpected molecular relatedness of IPF to neoplasm, diabetes, Alzheimer's disease, obesity, atherosclerosis, and arteriosclerosis. Furthermore, as a potential mechanism underlying this relatedness, we uncovered a putative molecular crosstalk system across the lung and the liver. In this inter-organ system, a secreted protein, kininogen 1, from hepatocytes in the liver interacts with its receptor, bradykinin receptor B1 in the lung. This ligand-receptor interaction across the liver and the lung leads to the activation of calmodulin pathways in the lung, leading to the activation of interleukin 6 and phosphoenolpyruvate carboxykinase 1 pathway across these organs. Importantly, we retrospectively identified several pre-clinical and clinical evidence supporting this inter-organ mechanism of IPF. In conclusion, such feedforward and feedback loop system across the lung and the liver provides a unique opportunity for the development of the treatment and/or diagnosis of IPF. Furthermore, the result illustrates a generative computational framework for machine-mediated synthesis of mechanisms that facilitates and complements the traditional experimental approaches in biomedical sciences.

Latent disease similarities and therapeutic repurposing possibilities uncovered by multi-modal generative topic modeling of human diseases.

Motivation: Human diseases are characterized by multiple features such as their pathophysiological, molecular and genetic changes. The rapid expansion of such multi-modal disease-omics space provides an opportunity to re-classify diverse human diseases and to uncover their latent molecular similarities, which could be exploited to repurpose a therapeutic-target for one disease to another. Results: Herein, we probe this underexplored space by soft-clustering 6955 human diseases by multi-modal generative topic modeling. Focusing on chronic kidney disease and myocardial infarction, two most life-threatening diseases, unveiled are their previously underrecognized molecular similarities to neoplasia and mental/neurological-disorders, and 69 repurposable therapeutic-targets for these diseases. Using an edit-distance-based pathway-classifier, we also find molecular pathways by which these targets could elicit their clinical effects. Importantly, for the 17 targets, the evidence for their therapeutic usefulness is retrospectively found in the pre-clinical and clinical space, illustrating the effectiveness of the method, and suggesting its broader applications across diverse human diseases. Availability and implementation: The code reported in this article is available at: https://github.com/skozawa170301ktx/MultiModalDiseaseModeling. Supplementary information: Supplementary data are available at Bioinformatics Advances online.