Drug target prediction through deep learning functional representation of gene signatures.

Many machine learning applications in bioinformatics currently rely on matching gene identities when analyzing input gene signatures and fail to take advantage of preexisting knowledge about gene functions. To further enable comparative analysis of OMICS datasets, including target deconvolution and mechanism of action studies, we develop an approach that represents gene signatures projected onto their biological functions, instead of their identities, similar to how the word2vec technique works in natural language processing. We develop the Functional Representation of Gene Signatures (FRoGS) approach by training a deep learning model and demonstrate that its application to the Broad Institute's L1000 datasets results in more effective compound-target predictions than models based on gene identities alone.

A Readily Scalable, Clinically Demonstrated, Antibiofouling Zwitterionic Surface Treatment for Implantable Medical Devices.

Unlike growth on tissue, microbes can grow freely on implantable devices with minimal immune system intervention and often form resilient biofilms that continuously pump out pathogenic cells. The efficacy of antibiotics used to treat infection is declining due to increased rates of pathogenic resistance. A simple, one-step zwitterionic surface modification is developed to significantly reduce protein and microbial adhesion to synthetic materials and demonstrate the successful modification of several clinically relevant materials, including recalcitrant materials such as elastomeric polydimethylsiloxane.

Recent advances in nanoengineering cellulose for cargo delivery.

The recent decade has witnessed a growing demand to substitute synthetic materials with naturally-derived platforms for minimizing their undesirable footprints in biomedicine, environment, and ecosystems. Among the natural materials, cellulose, the most abundant biopolymer in the world with key properties, such as biocompatibility, biorenewability, and sustainability has drawn significant attention. The hierarchical structure of cellulose fibers, one of the main constituents of plant cell walls, has been nanoengineered and broken down to nanoscale building blocks, providing an infrastructure for nanomedicine.

Exploiting Enzyme Alterations in Cancer for Drug Activation, Drug Delivery, and Nanotherapy.

Alterations of enzyme activity or enzyme levels in cancer tissue have been documented over the years. Taking advantage of these observations, various strategies for drug activation and drug delivery have been developed. One example is a mechanism called "prodrugs" that generates active drugs by enzyme cleavage.