publications
List of published and under-review research. (* indicates equal contributions.)
2025
- bioRxivRobust Prediction of Patient-Specific Cancer Hallmarks Using Neural Multi-Task Learning: a model development and validation studyShreyansh Priyadarshi, Camellia Mazumder, Bhavesh Neekhra, and 4 more authorsbioarXiv (preprint), 2025
Cancer progression is driven by a set of well-defined biological principles—collectively termed the “hallmarks of cancer”—yet current diagnostic approaches seldom incorporate these distinct molecular features into clinical practice. Despite substantial progress in molecular oncology, traditional methods like histopathological grading and immunohistochemical assays often fail to capture the complex interplay between cancer cells and the tumor microenvironment, emphasizing the need for robust computational frameworks capable of systematically quantifying hallmark-specific activity. Here, we address this gap by developing OncoMark, a high-throughput neural multi-task learning (N-MTL) framework designed to simultaneously quantify hallmark activities in tumor biopsies using transcriptomics data. We show that OncoMark achieves near-perfect accuracy, precision, recall, and F1 scores (>99%) in cross-validation, with external validation consistently exceeding 96.6% on five independent datasets. Further evaluation on eight additional datasets—including large-scale cancer cohorts (TCGA, MET500, CCLE, TARGET, PCAWG, POG570) and normal tissue datasets (GTEx, ANTE)—demonstrated high specificity for normal samples and robust sensitivity for hallmark prediction in cancer. By delivering a comprehensive and cost-effective molecular portrait of tumor biology and providing a user-friendly web platform accessible at [https://oncomark-ai.hf.space/](https://oncomark-ai.hf.space/), OncoMark has the potential to guide tailored treatment strategies and advance precision oncology. More broadly, this framework signifies a transformative step toward routine hallmark-based diagnostics, promising to improve patient outcomes by facilitating timely and precise tumor profiling.
@article{shreyansh2025oncomark, title = {Robust Prediction of Patient-Specific Cancer Hallmarks Using Neural Multi-Task Learning: a model development and validation study}, author = {Priyadarshi, Shreyansh and Mazumder, Camellia and Neekhra, Bhavesh and Biswas, Sayan and Chowdhury, Debojyoti and Gupta, Debayan and Haldar, Shubhasis}, year = {2025}, journal = {bioarXiv (preprint)}, doi = {https://doi.org/10.1101/2025.02.03.636380}, metatype = {Under Review} }
2024
- bioRxivComprehensive Enumeration of Cancer Stem-like Cell Heterogeneity Using Deep Neural NetworkDebojyoti Chowdhury*, Shreyansh Priyadarshi*, Sayan Biswas, and 3 more authorsbioarXiv (preprint), 2024
Cancer stem cells (CSCs), a distinct subpopulation within tumors, are pivotal in driving treatment resistance and tumor recurrence, posing substantial challenges to conventional therapeutic strategies. Precise quantification and profiling of these cells are essential for improving cancer treatment outcomes. We present ACSCeND, an advanced deep neural network model accompanied by a robust workflow, specifically developed to quantify cellular compositions from bulk RNA-seq data, enabling accurate CSC profiling. By integrating bulk RNA-seq data with insights derived from single-cell RNA-seq datasets, ACSCeND effectively captures the diversity and hierarchical organization of tumor-resident cell states, alongside cell-specific gene expression profiles (GEPs). Compared to current tissue deconvolution models, ACSCeND exhibits superior performance, achieving significantly higher Concordance Correlation Coefficient (CCC) values and lower Root Mean Square Error (RMSE) across various pseudobulk and real-world bulk tissue samples. Application of ACSCeND to TCGA and PRECOG datasets reveals a strong association between CSC abundance and poorer disease-free survival outcomes, underscoring the clinical relevance of CSCs in cancer progression. Furthermore, cell-specific GEPs for distinct CSC states unveil novel molecular signatures and illuminate the origins of CSC-driven tumor heterogeneity. In summary, ACSCeND provides a powerful, scalable platform for high-throughput quantification of cellular compositions and distinct potency states within normal tissues as well as highly heterogeneous tissues, such as tumors.
@article{shreyansh2024stem, title = {Comprehensive Enumeration of Cancer Stem-like Cell Heterogeneity Using Deep Neural Network}, author = {Chowdhury*, Debojyoti and Priyadarshi*, Shreyansh and Biswas, Sayan and Neekhra, Bhavesh and Gupta, Debayan and Haldar, Shubhasis}, year = {2024}, journal = {bioarXiv (preprint)}, doi = {https://doi.org/10.1101/2024.11.26.625418}, metatype = {Under Review} }
- NaturePan-cancer analyses suggest kindlin-associated global mechanochemical alterationsDebojyoti Chowdhury*, Ayush Mistry*, Debashruti Maity, and 5 more authorsNature Communications Biology, 2024
Kindlins serve as mechanosensitive adapters, transducing extracellular mechanical cues to intracellular biochemical signals and thus, their perturbations potentially lead to cancer progressions. Despite the kindlin involvement in tumor development, understanding their genetic and mechanochemical characteristics across different cancers remains elusive. Here, we thoroughly examined genetic alterations in kindlins across more than 10,000 patients with 33 cancer types. Our findings reveal cancer-specific alterations, particularly prevalent in advanced tumor stage and during metastatic onset. We observed a significant co-alteration between kindlins and mechanochemical proteome in various tumors through the activation of cancer-related pathways and adverse survival outcomes. Leveraging normal mode analysis, we predicted structural consequences of cancer-specific kindlin mutations, highlighting potential impacts on stability and downstream signaling pathways. Our study unraveled alterations in epithelial–mesenchymal transition markers associated with kindlin activity. This comprehensive analysis provides a resource for guiding future mechanistic investigations and therapeutic strategies targeting the roles of kindlins in cancer treatment.
@article{shreyansh2024kindlin, title = {Pan-cancer analyses suggest kindlin-associated global mechanochemical alterations}, author = {Chowdhury*, Debojyoti and Mistry*, Ayush and Maity, Debashruti and Bhatia, Riti and Priyadarshi, Shreyansh and Wadan, Simran and Chakraborty, Soham and Haldar, Shubhasis}, year = {2024}, journal = {Nature Communications Biology}, doi = {https://doi.org/10.1038/s42003-024-06044-5}, metatype = {Published} }
2023
- bioRxivNext-Gen Profiling of Tumor-resident Stem Cells using Machine learningDebojyoti Chowdhury*, Bhavesh Neekhra*, Shreyansh Priyadarshi, and 4 more authorsbioarXiv (preprint), 2023
Tumor-resident stem cells, also known as cancer stem cells (CSCs), constitute a subgroup within tumors, play a crucial role in fostering resistance to treatment and the recurrence of tumors, and pose significant challenges for conventional therapeutic methods. Existing approaches for identifying CSCs face notable hurdles related to scalability, reproducibility, and technical consistency across different cancer types due to the adaptable nature of CSCs. In this context, we introduce OSCORP, an innovative machine-learning-driven approach. This methodology quantifies and identifies CSCs, achieving almost 99% accuracy using biopsy bulk RNAseq data. OSCORP leverages genetic similarities between normal and cancer stem cells. By categorizing CSCs into four distinct yet dynamic potency states, this approach provides insights into the differentiation landscape of CSCs, unveiling previously undisclosed facets of tumor heterogeneity. In evaluations conducted on patient samples across 22 cancer types, OSCORP revealed clinical, transcriptomic, and immunological signatures associated with each CSC state. It has emerged as a comprehensive tool for understanding and addressing the complexities of cancer stem cells. Ultimately, OSCORP opens up new possibilities for more effective personalized cancer therapies and holds the potential to serve as a clinical tool for monitoring patient-specific CSC changes during treatment or follow-up care.
@article{shreyansh2023stem, title = {Next-Gen Profiling of Tumor-resident Stem Cells using Machine learning}, author = {Chowdhury*, Debojyoti and Neekhra*, Bhavesh and Priyadarshi, Shreyansh and Mukherjee, Swapnanil and Maity, Debashruti and Gupta, Debayan and Haldar, Shubhasis}, year = {2023}, journal = {bioarXiv (preprint)}, doi = {https://doi.org/10.1101/2023.11.24.568600}, metatype = {Under Revision} }
- bioRxivMethotrexate-modulated talin-dynamics drives cellular mechanical phenotypes via YAP signalingDebojyoti Chowdhury, Shukhamoy Dhabal, Madhu Bhatt, and 5 more authorsbioarXiv (preprint), 2023
Methotrexate is a well-known antineoplastic drug used to prevent cancer aggravation. Despite being a targeted therapeutic approach, its administration comes with the risk of cancer recurrence, plausibly through its proven off-target effect on focal adhesions. Since FA dynamics is dependent on force transmission through its constituent proteins, including talin, methotrexate might affect the mechanical activity of these proteins. Here we have combined single-molecule studies, computational dynamics, cell-based assays, and genomic analysis to unveil the focal adhesion-regulating role of methotrexate central to its effect on talin dynamics and downstream pathways. Interestingly, our single-molecule force spectroscopic study shows that methotrexate modulates the bimodal force distribution of talin in a concentration-dependent manner. Steered molecular dynamics reveal that methotrexate-talin interactions alter talin mechanical stability exposing their vinculin binding sites. Finally, we found that methotrexate-regulated talin-dynamics remodel cancer cell mechanical phenotypes like cell polarity, adhesion, and migration by regulating talin-vinculin association-mediated YAP signaling. These results further correlate with genomic analysis of methotrexate-treated patients, demonstrating its clinical importance. Taken together, these findings disseminate the effects of methotrexate-modulated mechanosensitivity of adhesion proteins on cellular events.
@article{shreyansh2023meth, title = {Methotrexate-modulated talin-dynamics drives cellular mechanical phenotypes via YAP signaling}, author = {Chowdhury, Debojyoti and Dhabal, Shukhamoy and Bhatt, Madhu and Maity, Debashruti and Chakraborty, Soham and Ahuja, Keshav Kant and Priyadarshi, Shreyansh and Haldar, Shubhasis}, year = {2023}, journal = {bioarXiv (preprint)}, doi = {https://doi.org/10.1101/2023.04.07.535979}, metatype = {Under Revision} }