International Business Machines Corporation

Zhaonan Sun

Jianying Hu

Ping Zhang

Ying Li

Fei Wang

Jing Li

Fei C Wang

Guo Tong Xie

Gang Hu

Haifeng Liu

Yu Cheng

Jingjing Tao

Soumya Ghosh

Zijun Yao

Zach Shahn

Liuyi Yao

Systems, computer-implemented methods, and computer program products to facilitate employing a probabilistic model to generate a continuous disease status index based on observational data are provided. According to an embodiment, a system can comprise a memory that stores computer executable components and a processor that executes the computer executable components stored in the memory. The computer executable components can comprise a model component that employs a probabilistic model to generate probability distributions of disease states of a disease of an entity based on observational data of the entity. The computer executable components can further comprise an index component that generates a disease status index of the disease based on the probability distributions of the disease states.

Generation of a disease status index using a probabilistic model and observational data

Techniques are described that facilitate determining potential cancer gene therapy targets by joint modeling of cancer survival events. In one embodiment, a computer-implemented comprises employing, by a device operatively coupled to a processor, a multi-task learning model to determine active genetic factors respectively associated with different types of cancer based on cancer survival data and patient genomic data for groups of patients that respectively survived the different types of cancer. The computer-implemented method further comprises, determining, by the device, common active genetic factors of the active genetic factors that are shared between two or more types of cancer of the different types of cancer.

Determining potential cancer therapeutic targets by joint modeling of survival events

A system (or method) for generation and employment of disease progression model(s) that facilitates identifying and indexing discriminative features for disease progression in observational data. The disease progression prediction system comprises a processor that executes computer executable components stored in memory. A receiving component receives and learns observational patient data. A model generation component builds a preliminary disease progression model. An identification component identifies discriminative clinical features for different disease stages. A ranking component ranks discriminative powers of clinical features for respective pairs of disease stages; wherein the model generation component employs the ranked features to generate a final disease progression model.

Identifying and indexing discriminative features for disease progression in observational data

Predicting beneficial drug combinations mitigating adverse drug reactions identifies drug combinations and associated target adverse drug reaction from a spontaneous reporting system containing case reports of drugs and associated adverse drug reactions. Each drug combination comprises a first drug and a second drug, and a propensity score is computed for each drug in each group. This propensity score expresses a probability of being exposed to a given drug based on other co-prescribed drugs and reported indications, which reflect patient characteristics. Associations are computed for each drug as well as drug interaction. Among the associations, the sum of the associations of the second drug and the interaction effect represents the predicted beneficial score expressing whether the second drug alters the chance of developing the target adverse drug reaction for patients on the first drug. The interaction effect is referred to as predicted interaction score, and represents antagonistic or synergistic drug interactions.

Data-driven prediction of drug combinations that mitigate adverse drug reactions

An embodiment of the invention receives input including a list of drugs, drug characteristics of each drug, and known drug-disease associations including a disease and a drug having a threshold efficacy for treating the disease. For each drug in the list of drugs, a processor predicts whether the drug meets a threshold efficacy for treating a first disease based on the drug characteristics and the drug-disease associations. For each drug in the list of drugs, the processor predicts whether the drug meets a threshold efficacy for treating a second disease based on the drug characteristics and the predicting of whether the drug meets the threshold efficacy for treating the first disease. Output is generated output based on the predictions, the output including an identified drug-disease association, an identified disease-disease association, an identified chemical fingerprint for the first disease, and an identified chemical fingerprint for the second disease.

Evidence boosting in rational drug design and indication expansion by leveraging disease association

A method, system, and computer program product, include obtaining a plurality of training sets of samples and a plurality of testing sets of samples from a data set of clinical records, the plurality of training sets of samples and the corresponding plurality of testing sets of samples including different samples of the data set, determining statistical significance for a plurality of risk factors of a model from the plurality of training sets of samples, determining a plurality of performance metrics of the model from the plurality of testing sets of samples, and determining an evaluation metric of the plurality of risk factors for the model based on the statistical significance and the plurality of performance metrics, the evaluation metric being for improving accuracy of the model.

System, method and computer program product for evaluation and identification of risk factor

An apparatus, method and computer program product for proactive comprehensive generic risk screening. The method performs proactive comprehensive generic risk screening by implementing steps of training comprising steps of receiving cross domain risks and features, optimizing linkage regularization using the received features and the received cross domain risks, said linkage regularization comprising multi-task predictive model training, feature selection and ranking, risk association learning and risk association selection, and outputting patient risk scores, identified high risk patients, risk factors for risks and risk groups, and risk groups and risk associations and calculating risk score for an individual patient comprising steps of receiving individual features comprising patient information, performing said linkage regularization using the received individual features and outputting patient risk scores for said individual patient, and high risk for said individual patient. The calculating risk score can be performed for more than one patient.

Growing community of inventors

Elmsford, NY, United States of America

Zhaonan Sun