Balachandran Manavalan

Phosphorylation, a critical post-translational modification, is extensively altered during viral infections, including SARS-CoV-2, where it plays a central role in modulating host–pathogen interactions. Accurately identifying these specific phosphorylation sites is crucial for understanding viral pathogenesis, prioritizing antiviral targets, guiding therapeutic strategies, and strengthening preparedness for future viral outbreaks. Although several computational tools have been proposed to complement experimental phosphoproteomics, existing methods often show limited robustness, cross-viral generalizability, and interpretability. To address these challenges, we developed PhoSARte, an interpretable computational framework that integrates Siamese network-based contrastive learning (SCL) with pretrained protein language models (PLMs) to accurately identify phosphorylation sites in SARS-CoV-2-infected cells. PhoSARte employs a unique dual-stream architecture: PLMs capture contextual protein sequence representations, while the SCL module, comprising a transformer-based encoder and an attention-based bidirectional gated recurrent unit, learns discriminative and similarity-preserving representations of protein sequence pairs using a contrastive loss function. The integration of these complementary representations substantially improves the robustness and generalizability of the framework. PhoSARte was rigorously benchmarked on phosphoproteomics datasets derived from infected A549 (Homo sapiens) and Vero E6 (Chlorocebus sabaeus) cells, as well as their combined dataset. Through rigorous cross-cell-type validation and testing, PhoSARte demonstrated superior performance, significantly outperforming current state-of-the-art methods. Importantly, an external cross-viral case study on entirely unseen adenovirus type 2-infected human IMR-90 cells confirmed the broad transferability of PhoSARte, demonstrating its capacity to generate actionable biological hypotheses under novel viral stress conditions. Furthermore, advancing beyond traditional black-box predictors, PhoSARte integrates an in silico mutagenesis analysis that decodes complex deep learning embeddings to successfully extract biologically relevant motif signatures. PhoSARte is freely accessible at https://balalab-skku.org/PhoSARte/, providing an accessible, interpretable, and adaptable framework for virus-associated phosphorylation site prediction, antiviral target prioritization, and host-directed therapeutic discovery.

Accurate pathway activity inference from single-cell RNA sequencing (scRNA-seq) data is hindered by sparsity, technical noise, and the weak yet coordinated nature of transcriptional programs. Existing methods typically aggregate expression values over predefined gene sets, which can obscure context-dependent regulatory structure. Here, we present Graph-based Pathway Activity Scoring (GraphPAS), a hierarchical graph learning framework for recovering coherent pathway-level structure from scRNA-seq data. Systematic benchmarking across scRNA-seq datasets showed that GraphPAS consistently achieved higher adjusted Rand index, normalized mutual information, and silhouette width than AUCell and scapGNN, while maintaining greater robustness under dropout and Gaussian noise perturbations. Applied to adenomyosis scRNA-seq data, GraphPAS revealed enrichment of programmed cell death programs in macrophages. Pain-associated samples showed elevated apoptosis, ferroptosis, and necroptosis signatures accompanied by inflammatory activation, implicating macrophage-centered cell death remodeling in the adenomyosis microenvironment.

Interleukin-6 (IL-6) is a key immunomodulatory cytokine implicated in diverse physiological processes and pathological conditions, including autoimmune diseases, cancers, and cytokine storms. Immunogenic peptides capable of inducing IL-6 expression are key modulators of host immune responses and represent promising candidates for therapeutic design and epitope-based vaccine development. However, experimental identification of IL-6-inducing peptides remains laborious and unsuitable for large-scale screening. Although existing computational approaches show promise, many often struggle to capture both global contextual semantics and local motif-level features essential for peptide immunogenicity. To address these limitations, we present CONTRA-IL6, a novel deep learning framework that integrates Transformer fusion and convolutional localization modules with stacked pretrained protein language model embeddings to predict IL-6-inducing peptides. Comprehensive benchmarking on an independent dataset demonstrates that CONTRA-IL6 achieves superior predictive performance over six state-of-the-art predictors. Notably, it achieves the highest Matthews correlation coefficient (MCC, 0.504) and F1 (0.549) and improves over the best-performing existing method by 3.2% in MCC and 4.3% in F1, demonstrating balanced and robust performance. Feature space visualizations (uniform manifold approximation and projection, kernel density estimation) showed clear class separation, while 1D gradient-weighted class activation mapping++ highlighted strong attention to specific C-terminal regions. Crucially, we moved beyond these attribution methods by employing in silico mutagenesis, which causally confirmed the functional importance and physicochemical constraints. Ablation studies further confirmed the synergistic contribution of global and local modules to model performance. CONTRA-IL6 offers a robust, scalable, and interpretable solution for immunoinformatics research. The standalone package is freely available at https://pypi.org/project/contra-il6/ to facilitate broader community use.

Bitter peptides (BPs), derived from the hydrolysis of proteins in food, play a crucial role in both food science and biomedicine by influencing taste perception and participating in various physiological processes. Accurate identification of BPs is essential for understanding food quality and potential health impacts. Traditional machine learning approaches for BP identification have relied on conventional feature descriptors, achieving moderate success but struggling with the complexities of biological sequence data. Recent advances utilizing protein language model embedding and meta-learning approaches have improved the accuracy, but frequently neglect the molecular representations of peptides and lack interpretability. In this study, we propose xBitterT5, a novel multimodal and interpretable framework for BP identification that integrates pretrained transformer-based embeddings from BioT5+ with the combination of peptide sequence and its SELFIES molecular representation. Specifically, incorporating both peptide sequences and their molecular strings, xBitterT5 demonstrates superior performance compared to previous methods on the same benchmark datasets. Importantly, the model provides residue-level interpretability, highlighting chemically meaningful substructures that significantly contribute to its bitterness, thus offering mechanistic insights beyond black-box predictions. A user-friendly web server ( https://balalab-skku.org/xBitterT5/ ) and a standalone version ( https://github.com/cbbl-skku-org/xBitterT5/ ) are freely available to support both computational biologists and experimental researchers in peptide-based food and biomedicine. We propose xBitterT5, a novel multimodal transformer-based framework for the identification of BPs. By utilizing the pretrained BioT5+ model, xBitterT5 effectively extracts high-level representations from both the peptide sequences and their corresponding SELFIES molecular representation. This dual-modality approach enables a more comprehensive understanding of the peptide sequence by leveraging its molecular string, leading to substantial improvements in performance across two benchmark datasets. Additionally, xBitterT5 offers interpretability by identifying key molecular substructures that contribute to bitterness, thereby providing mechanistic insights essential for peptide-based food and drug applications.