Network-Based Drug-Target Interaction Prediction with Probabilistic Soft Logic

Abstract

Drug-target interaction studies are important because they can predict drugs' unexpected therapeutic or adverse side effects. In silico predictions of potential interactions are valuable and can focus effort on in vitro experiments. We propose a prediction framework that represents the problem using a bipartite graph of drug-target interactions augmented with drug-drug and target-target similarity measures and makes predictions using probabilistic soft logic (PSL). Using probabilistic rules in PSL, we predict interactions with models based on triad and tetrad structures. We apply (blocking) techniques that make link prediction in PSL more efficient for drug-target interaction prediction. We then perform extensive experimental studies to highlight different aspects of the model and the domain, first comparing the models with different structures and then measuring the effect of the proposed blocking on the prediction performance and efficiency. We demonstrate the importance of rule weight learning in the proposed PSL model and then show that PSL can effectively make use of a variety of similarity measures. We perform an experiment to validate the importance of collective inference and using multiple similarity measures for accurate predictions in contrast to non-collective and single similarity assumptions. Finally, we illustrate that our PSL model achieves state-of-the-art performance with simple, interpretable rules and evaluate our novel predictions using online data sets.

A Collective, Probabilistic Approach to Schema Mapping Using Diverse Noisy Evidence

“…More importantly, in contrast to Markov Logic, PSL avoids the hard combinatorial optimization problem and instead provides scalable inference with guarantees on solution quality. This advantage has proven crucial also for applications of PSL in knowledge graph identification [15] and data fusion [16], [17].…”

“…Using these notions, we define our probabilistic optimization problem using probabilistic soft logic (PSL) [14], a scalable probabilistic programming language based on weighted logical rules. PSL has been used successfully for a variety of data and knowledge integration problems, including knowledge graph identification [15] and data fusion [16], [17]. It however did not support the kind of open world reasoning required for mapping selection, where we need to express constraints over the existence of elements in a set satisfying certain conditions, namely, st tgds in the mapping explaining tuples in the data example, and furthermore, preferences over these elements are available.…”

Heter-LP: A heterogeneous label propagation algorithm and its application in drug repositioning

“…Osteoarthritis with mild chondrodysplasia is a type of skeletal disease due to the mutation of type II procollagen (COL2A1). It causes a progressive degeneration of the articular cartilage of joints with mild spinal chondrodysplasia 19,20 . 19 "Chondrodysplasia is a heterogeneous group of bone dysplasias, the common characteristic of which is stippling of the epiphyses in infancy."…”

Machine learning prediction of oncology drug targets based on protein and network properties

“…Many data-driven approaches have focused on gene expression changes after drug treatment to predict similarity between drugs and potentially predict shared targets [7][8][9]. Several studies combined different data types to improve the prediction of drug similarity [10] and to predict drug-target interactions [11,12]. Other methods focused on predicting the probability of success in a clinical trial by estimating the toxicity based on several chemical properties, drug-likeness measures of the molecules and the target properties [13].…”