Our Article about a new integration pipeline which combines multi-omics, clinical, and lab data to achieve up to 96% accuracy in diagnosing autoimmune diseases, offering adaptable applications for broader disease diagnostics.

Automated translation of narrative care guidelines into structured components for integration into electronic clinical decision support systems - a feasibility study

The paper investigates the dengue antibody response using advanced sequencing and machine learning, revealing insights into diversity changes, repertoire architecture, and identifying rare broadly neutralizing antibody sequences.

Em Care is a fully open-source, FHIR-native solution that can be configured to meet country- or site-specific requirements and integrated to Member State’s existing digital solutions. It is available for Android devices and built using the Android FHIR SDK which supports HL7 FHIR®/CQL. It will be available from the Google Play store or as a direct download to an Android Package Kit (APK)

An applied computational intelligence project to determine whether LightGBM is a suitable replacement for the random forest classifier widely used in QSAR models. Published and presented at the IJCACI 2022 conference.

A curated Database of immune repertoires and metadata in the AIRR format annotated using IgBLAST.

A full-stack application to search and filter for desired genes that have been downloaded from the NCBI FTP server using Java Spring Boot, Maven, Apache Tomcat, SQL, npm and React.

Predicting the bioactivity of small molecules in SMILES format towards a known target molecule Monoamine Oxidase B (MAO-B), an enzyme for the inhibition of dopamine and therefore a known drug target for Parkinson’s disease.

The paper explores the use of high-throughput sequencing and machine learning techniques, including neural networks, random forests, and support vector machines, to analyze the immune response to dengue virus at both the antibody repertoire and sequence levels, identifying detectable signatures that could contribute to antibody discovery and vaccine design.

PERSONALIS, a clinical decision support system aimed to demultiplex complex autoimmune diseases by applying machine learning models to various types of medical and molecular datasets for the diagnosis and treatment.

Microsoft Hololens application written in C# using Unity and Vuforia for drug package recognition, which can present the associated ingredients and their corresponding molecules in 3D. Mobile solution for Android as well as iOS to enable a playful interaction of pharmaceutical products.

Development of a hospital architecture for the analysis of image data from medical diagnostic images. Full-stack solution with the ability to filter results according to specific criteria or by using tags. The app was built using TypeScript, Vue.js, Node, Docker & PACS Server.