Abstract The pharmaceutical industry is experiencing a profound transformation driven by rapid advancements in artificial intelligence (AI), increasing research and development complexity, and the expiration of key patents. Agentic-AI systems, which enable autonomous reasoning, decision-making, and adaptive learning beyond traditional data-driven automation, represent a paradigm shift in this context. This review examines the impact of Agentic-AI across the pharmaceutical value chain, including supply chain orchestration, manufacturing optimization, early-stage drug discovery, process development, and regulatory intelligence. These systems enhance operational resilience, reduce development timelines, and accelerate the identification of new therapeutic candidates by fostering dynamic human-AI collaboration. However, the widespread adoption of Agentic-AI introduces significant challenges related to workforce integration, ethical accountability, governance, and regulatory compliance. Addressing these challenges is essential for reliable and scalable deployment. This study identifies key factors shaping Agentic-AI’s influence over the next decade and provides a forward-looking perspective on its evolving role in pharmaceutical innovation.
Abstract Optimizing protein expression and buffer conditions is a major challenge in life-science research and biotechnology. We present an LLM-powered extraction pipeline that systematically converts unstructured scientific literature into a structured, queryable database. Leveraging advanced optical character recognition, custom reranker models, and large language models, our system achieves ~97% extraction accuracy. By processing 100 353 structural biology publications associated with the RCSB Protein Data Bank, we constructed a database containing 947 153 distinct buffer entries and 73 210 expression protocols for 35 690 unique proteins, establishing the largest resource of its kind. The ProBufEx (https://www.probufex.org/) web platform enables rapid exploration of optimal conditions by protein properties or functions. Case studies reveal distinct patterns in chemical usage on proteins with different post-translational modifications, and function-dependent expression strategies for proteins with different enzymatic activities. This resource transforms vast structural biology literature into actionable knowledge, accelerating evidence-based protocol design, reducing trial-and-error, and streamlining experiments in protein science.
Abstract Body fluid identification methods have become increasingly well-established, but deconvoluting multi-component mixtures remains a significant challenge. When dealing with complex mixtures, methods based on mRNA coding-region SNPs (cSNPs) still face limitations such as limited polymorphisms of markers and insufficient population specificity. mRNA coding-region microhaplotypes (cMHs), combining tissue specificity with high discrimination power, offer a promising alternative for donor attribution in body fluid stains. In this study, we developed a targeted RNA sequencing panel for East Asian populations, comprising 36 body fluid-specific genes containing 48 cMHs (including 108 cSNPs), along with five genes lacking cMHs. This panel enables both tissue source identification and donor attribution for five types of body fluids and their mixtures. Through specificity evaluation of single-source body fluid stains, we identified six novel body fluid-specific genes. Furthermore, by integrating machine learning algorithms with an in silico mixing strategy, we developed a body fluid classifier that achieved accurate inference for both single-source fluids and complex mixtures. The panel's cumulative discrimination power across the five body fluid types ranged from 0.99568126 (peripheral blood) to 0.99999999 (vaginal secretion). In conclusion, the dual-purpose panel demonstrates robust performance in both tissue source identification and donor attribution for single-source body fluids and complex mixtures.
Abstract Spatial transcriptomics (ST) facilitates the spatial resolution of gene expression profiling; however, many widely used spot-based sequencing platforms capture transcripts from multiple neighboring cells at each location, producing mixed profiles that require accurate cell-type deconvolution. The wavelet-denoised graph attention framework for spatial transcriptomics deconvolution (WDGTA), a graph attention-based framework for ST cell-type deconvolution, was developed and validated on five datasets from two species and four organs. Through the integration of real with pseudo-ST spots, WDGTA introduces a wavelet-based multiscale denoising and reconstruction module, constructs a weighted topological graph using random projection forests, and applies multi-head graph attention to infer cell-type proportions in spatial spots. WDGTA achieved optimal performance in comparison with seven established deconvolution methods. Ablation and discrete wavelet transform (DWT)-specific analyses confirmed the contributions of wavelet denoising, weighted graph construction, and semi-supervised learning. Furthermore, DWT performance remained stable after feature-order perturbation. In mouse brain, cerebellum and ultraviolet radiation (UVR) skin data, WDGTA recovered biologically consistent spatial cell-type patterns and demonstrated robust cross-platform applicability and improved resolution of low-abundance melanocyte signals. Furthermore, it accurately captured melanocyte-associated spatial changes under conditions of UVR exposure and drug intervention. WDGTA provides a unified framework for ST deconvolution across the evaluated tissue contexts. In UVR-exposed skin, it resolved localized melanocyte-associated patterns that were concordant with marker-gene evidence and detected changes associated with UVR exposure and drug treatment. These findings support its use for characterizing spatial cell-type organization and low-abundance cell populations in heterogeneous tissues.
Abstract Ribosome heterogeneity is increasingly recognized as a key factor in various biological contexts, including development, cancer, and immune responses. To enable systematic exploration of this diversity, we previously developed a bioinformatics method that utilized rRNA reads from Ribo-seq data, detecting changes in ribosomal protein composition (dripARF) and ribosome collisions (dricARF) between conditions simply and efficiently. The method works by projecting differences in rRNA read abundance onto the 3D-structure of the ribosome, following which we perform enrichment tests on predefined heterogeneity-specific rRNA position sets. In this study, we apply this tool to publicly available Ribo-seq datasets to understand the impact of experimental protocol variations and to identify shared biological signals across studies. Our analyses show that the rRNA regions assigned to our prediction sets are consistently enriched in rRNA reads, supporting the core assumptions of our approach. We further examine how various protocol factors influence rRNA read distribution and implement a single-sample rRNA position set enrichment analysis-based weight to limit false positives. Finally, we highlight consistent ribosomal changes across these studies, including prominent co-predictors such as ribosome collisions, eS25, and uS19.
