Developed biomarker discovery methods tailored to various learning paradigms, focusing on human aging and cellular senescence.
Used a positive-unlabeled learning approach to detect rare aging cell populations that led to a list of genes potentially enriched in aged (senescent) cells.
Developed a transformer-based neural network for temporal graph structure learning. The method provided insights into the evolution of gene networks with advancing age.
Selected Publications
Integrating patients in time series clinical transcriptomics data. Hasanaj et al., Intelligent Systems for Molecular Biology (ISMB) (2024)
Multiset multicover methods for discriminative marker selection. Hasanaj et al., Cell Reports Methods (2022)
Interactive single-cell data analysis using Cellar.Hasanaj et al., Nature Communications (2022)
Competitions
4th place, NeurIPS AutoML Decathlon Competition
Silver Medal, International Mathematics Competition for University Students (IMC)
Bronze Medal (team), ACM, Southeastern Europe Regional Programming Contest
2nd Place (team), National Programming Contest, Bulgaria
Honorable Mention, International Mathematical Olympiad (IMO)