A5003971810- vaccines and immunoinformatics approaches
- RNA and protein synthesis mechanisms
- Machine Learning in Bioinformatics
- T-cell and B-cell Immunology
- RNA Research and Splicing
- Bioinformatics and Genomic Networks
- Single-cell and spatial transcriptomics
- Biochemical and Structural Characterization
- CAR-T cell therapy research
- Bacterial Genetics and Biotechnology
- Caveolin-1 and cellular processes
- Computational Drug Discovery Methods
- Immune Cell Function and Interaction
- Pancreatic function and diabetes
University of Pittsburgh
2023-2025
Chinese University of Hong Kong
2021
Abstract The explosion of sequence data has allowed the rapid growth protein language models (pLMs). pLMs have now been employed in many frameworks including variant-effect and peptide-specificity prediction. Traditionally, for protein-protein or peptide-protein interactions (PPIs), corresponding sequences are either co-embedded followed by post-hoc integration concatenated prior to embedding. Interestingly, no method utilizes a representation interaction itself. We developed an LM (iLM),...
We introduce a method to generate synthetic protein sequences which are predicted be resistant certain antibiotics. did this using 6,023 genes that were antibiotics in the intestinal region of human gut and fed as input Wasserstein generative adversarial network (W-GAN) model variant original has been known perform efficiently when it comes mimicking distribution real data order new is similar style was training
Abstract Background The alpha, beta and gamma clustered protocadherins (cPcdhs) genes encode homotypic cell adhesion molecules that are highly expressed in the brain. They well described to be a combinatorial fashion specify neuronal identity for coding synaptic connectivity. Unexpectedly, we recently uncovered evidence from studies mice human showing PCHDA variants can cause congenital heart defects. As all of alpha or PCDH cluster share common 3' end homologous, expression these gene...
Immunomodulatory variants that lead to the loss or gain of specific protein interactions often manifest only as organismal phenotypes in infectious disease. Here, we propose a network-based approach integrate genetic variation with structurally resolved human interactome network prioritize immunomodulatory COVID-19. We find that, addition pass genome-wide significance thresholds, at interface protein-protein interactions, even though they do not meet are equally immunomodulatory. The...