Dr. Badhan Das
Assistant Professor - Full Time
badhan.das@bracu.ac.bd
+8801346068309
Websites
https://badhan023.github.io/Address
CSE Department
4th floor, Room No # 4G19,
Brac University,
Kha 224 Bir Uttam Rafiqul Islam Avenue,
Merul Badda, Dhaka, Bangladesh
Badhan Das is an Assistant Professor in the Department of Computer Science and Engineering at BRAC University and a former Postdoctoral Associate in the Department of Biology at New York University's Center for Genomics and Systems Biology. She earned her Ph.D. and M.S. in Computer Science from Virginia Tech (USA), following a B.Sc. in Computer Science from the Bangladesh University of Engineering and Technology (BUET). With an extensive background in both university instruction and interdisciplinary research, Badhan’s expertise spans Computational Biology, Graph Theory, and Machine Learning applications in Genomics. Her peer-reviewed work includes innovative contributions to viral evolution modeling, SARS-CoV-2 network dynamics, structural protein classification, and metagenomic surveillance tools.
At the intersection of machine learning, graph theory, and advanced genomics, Badhan’s current research program focuses on decoding the evolutionary and regulatory complexities of plant and human genomes. Her lab actively investigates deep learning frameworks—including CNNs, Transformers, and Graph Neural Networks (GNNs)—for TF-DNA binding site prediction. A key area of exploration within her group is applying transfer learning to translate regulatory models from well-characterized plant models, such as Arabidopsis, to complex agricultural crops like maize. Additionally, her research deeply explores the effectiveness of pangenome graphs as an alternative to traditional linear reference genomes, applying Genomic Large Language Models (LLMs) and graph-aware tokenization to capture structural variations and population-scale genetic diversity.
Das, B. and Heath, L.S., 2026. ViraFit: Tunable Fitness Model for Viral Evolution Within a Contact Network. Journal of Computational Biology, p.15578666261444396.
Das, B. and Heath, L.S., 2025. Variant evolution graph: Can we infer how SARS-CoV-2 variants are evolving?. Plos one, 20(6), p.e0323970.
Emon, M.I., Das, B., Thukkaraju, A.R. and Zhang, L., 2024, November. DeePSP-GIN: identification and classification of phage structural proteins using predicted protein structure, pretrained protein language model, and graph isomorphism network. In Proceedings of the 15th ACM International Conference on Bioinformatics, Computational Biology and Health Informatics (pp. 1-6).
Das, B., Emon, M.I., Moumi, N.A., Sein, J., Pruden, A., Heath, L.S. and Zhang, L., 2022. HT-ARGfinder: a comprehensive pipeline for identifying horizontally transferred antibiotic resistance genes and directionality in metagenomic sequencing data. Frontiers in Environmental Science, 10, p.901917.
Bardhan, R., Rabeya, M., Mahmood, M., Das, B. and Bhuiyan, H., 2021, August. An Area-efficient 2–to–4 Decoder Design Based on Quantum Dot Cellular Automata. In 2021 International Conference on Science & Contemporary Technologies (ICSCT) (pp. 1-5). IEEE.
Moumi, N.A., Das, B., Tasnim Promi, Z., Bristy, N.A. and Bayzid, M.S., 2019. Quartet-based inference of cell differentiation trees from chip-seq histone modification data. Plos one, 14(9), p.e0221270.
Rabeya, M., Mahmood, M., Das, B., Bardhan, R. and Tareque, M.H., 2019, February. An efficient design of 4-to-2 encoder and priority encoder based on 3-dot qca architecture. In 2019 International conference on electrical, computer and communication engineering (ECCE) (pp. 1-6). IEEE.
Das, B., Mahmood, M., Rabeya, M. and Bardhan, R., 2019, January. An effective design of 2: 1 multiplexer and 1: 2 demultiplexer using 3-dot QCA architecture. In 2019 International Conference on Robotics, Electrical and Signal Processing Techniques (ICREST) (pp. 570-575). IEEE.
Accepting
As: |
|
Level: |
Undergraduate & Postgraduate |
Type: |
|
My research lies at the intersection of Computational Biology, Graph Theory, and Machine Learning. I am highly interested in mentoring motivated undergraduate and graduate thesis students seeking to develop cutting-edge computational solutions to complex biological data. My lab currently focuses on leveraging deep learning and next-generation graph structures to decode genome regulation and population-scale genetic diversity.
Gene expression is tightly regulated by Transcription Factors (TFs) binding to specific DNA sequences, a process critical to understanding both human diseases and crop resilience. This research project focuses on developing advanced deep learning models (such as CNNs and Transformers) to predict TF-DNA binding sites in human and plant genomes. A core challenge we will tackle is transfer learning: training robust predictive models on the well-characterized Arabidopsis genome and adapting them to map regulatory mechanisms in the larger, more complex maize genome. Students joining this project will gain hands-on experience in cutting-edge genomic deep learning, cross-species model generalization, and plant regulatory networks.
Traditional genomic pipelines rely heavily on a single, linear reference genome, which introduces a severe "reference bias" by omitting structural variants, insertions, and diverse haplotypes across a population. Pangenome graphs resolve this limitation by modeling an entire species' genetic diversity as an interconnected network. This project aims to explore the effectiveness of pangenome graphs across various machine learning solutions in computational biology, where linear reference genomes are typically used. Students will investigate the integration of Graph Neural Networks (GNNs) and Genomic Large Language Models (LLMs) into graph topologies. We will tackle key challenges such as designing graph-aware tokenization techniques for genomic Transformers and adapting downstream prediction tasks, including variant calling and regulatory sequence analysis, directly onto native graph data structures. This project is ideal for students eager to work on advanced graph structures, machine learning optimization, and next-generation bioinformatics.
