Maryam Haghani

Maryam Haghani

Ph.D. student 👩🏻‍🎓 Bioinformatics, Applied Machine Learning, Computational Biology

My Research

Welcome to my research page. I am a Ph.D. student in Computer Science specializing in Computational Biology.

Research Interests

  • Computational Biology and Bioinformatics
  • Algorithm Design and Analysis
  • Applied Deep Learning and NLP
  • Large Language Models
  • Data Science for Bioinformatics

Projects

Cross-Species TFBS Prediction in Plant Genomes using DNA Foundation Models (2025)

Haghani, M.; Dhulipalla, K. V.; Li, S. (Under review)

  • Fine-tuning and benchmarking three large pretrained DNA foundation models (DNABERT-2, AgroNT, HyenaDNA) on DAP-seq data for ABF transcription factors in Arabidopsis thaliana and Sisymbrium irio.
  • Evaluated performance across cross-chromosome, cross-dataset, and cross-species protocols.
  • Demonstrated that HyenaDNA achieves near–state-of-the-art accuracy with over 10× faster training time, enabling scalable, genome-wide TFBS prediction in plants.

Analysis of Arabidopsis Nuclear Envelope Proteins using scRNA-Seq Data (2025)

  • Performed scRNA-seq analysis of Arabidopsis nuclear envelope proteins from root cells using R.
  • Identified cell type-specific genes and analyzed co-expression patterns.
  • Generated gene clusters and constructed a co-expression network with hdWGCNA.
  • Visualized co-expression networks to highlight functional links between genes.

HostVirusPair – Enhancing Host-Viral Protein Complex Prediction (2024)

Haghani, M.; Bhattacharya, D.; Murali, T. M. (Under submission)

  • Developed a novel MSA pairing algorithm using sequence-based deep learning.
  • Integrated interchain coevolutionary signals to improve AlphaFold-Multimer predictions.
  • Achieved higher DockQ scores and enhanced accuracy in structure predictions.

NEFFy – A Toolbox for NEFF Calculation and MSA Conversion (2025)

Haghani, M.; Bhattacharya, D.; Murali, T. M. Bioinformatics, 2025 – Website

  • Developed a tool for calculating Number of Effective Sequences (NEFF).
  • Enabled MSA format conversion across protein, DNA, and RNA alphabets.
  • Implemented column-wise NEFF calculation and multimeric MSA handling.
  • Built in C++ with a Python library for integration into workflows.