Software
Our team excels in crafting cutting-edge statistical software packages and interactive web applications tailored for data analysis and visualization in research settings. Here’s a curated list of our notable creations:
EHR-based Phenotyping Tools
KOMAP: A web-based platform for knowledge-driven online multimodal automated phenotyping using EHR data.
https://doi.org/10.1101/2023.09.29.23296239MAP: An R package enhancing patient phenotyping accuracy and efficiency through EHR-biorepository integration.
https://doi.org/10.1093/jamia/ocz066PheCAP: An R package enabling high-throughput phenotyping via a standardized automated pipeline, leveraging SAFE and machine learning for model development and validation.
https://doi.org/10.1038/s41596-019-0227-6PheNorm: An R package offering unsupervised phenotyping for EHR data, eliminating the need for expert-labeled samples.
https://doi.org/10.1093/jamia/ocx111
Knowledge Graphs and Databases
ARCH: A sophisticated tool for analyzing and visualizing EHR data relationships.
https://doi.org/10.1101/2023.05.14.23289955Hierarchies: A comprehensive toolkit for exploring medical code hierarchies used in standardizing health information across various healthcare contexts.
ONCE: A feature-generation search engine identifying related narrative and codified EHR features based on target diseases.
https://doi.org/10.1101/2023.09.29.23296239
Risk Prediction Frameworks
SCANER: An R package for estimating survival curves using both labeled and unlabeled data within a semi-supervised learning framework.
https://doi.org/10.1016/j.jbi.2023.104425SRAT: A statistical tool for assessing genetic variant-outcome associations, accounting for familial correlations and covariate adjustments.
https://doi.org/10.1111/biom.12643