Automatic Inference of Smart Data Discovery Interfaces for Rare Disease Datasets

Café Variome is a flexible, web-based, data-discovery tool that offers a query language for checking the existence of biomedical data held in a federation of heterogenous data sources of rare diseases. Currently, there is a growing need for user-friendly graphical interfaces that are web-based to assist bioinformatics researchers look for cohorts of patients, across widely heterogenous data sources. Moreover, data owners who are usually clinicians, hospitals, or local trusts, lack the technical programming skills to create an interface that can help their data be queried and hence discovered. In this work, we present the design of VForms, a platform for automatically inferring smart user interfaces (UIs) from heterogeneous datasets about rare diseases. This platform consists of a domain-specific modeling language (DSL) for specifying UIs for genomic datasets. A VForm is realized as a ReactJS web form that allows rare disease medical researchers define optimized queries over rare disease datasets. Hence each VForm is a domain-specific form-based query language. VForms infers the conceptual model from a given dataset, from which a domain-specific query UI in ReactJS is automatically generated using model-to-model transformations in YAMTL. The generation process is parameterized so that medical researchers can customize the generation of UIs. The goal of this work is to demonstrate the potential of using model-driven engineering (MDE) to improve the development of UIs for genomic datasets of patients with rare diseases. By using VForms, data owners who lack the technical programming skills to create a UI that can help their data be queried and discovered, will be able to do so with minimal effort and in a more user-friendly way. Additionally, VForms allows for quick adaptation of the implementation of query interfaces to new user requirements.