Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing最新文献

{"title":"Leveraging Generative AI for Interpretable Clinical Decision Making Through Causal Graphs.","authors":"Mehmet Eren Ahsen, Rand Kittani, Travis Gerke, Laya Krishnan, Sean Rogan, Erick R Scott","doi":"10.1142/9789819824755_0022","DOIUrl":"10.1142/9789819824755_0022","url":null,"abstract":"<p><p>Clinical AI systems' lack of interpretability limits their adoption in evidence-based medicine. To address this challenge, we propose a computational framework that harnesses generative AI's medical knowledge to create interpretable structural causal models (SCMs) for clinical decision support, quality improvement evaluation, and population health management. We evaluated our approach through a case study using data from the Midwest Healthcare Conference Causal Diagram Challenge, where we compared transformer-based large language models against human performance on a complex causal reasoning task: estimating COVID- 19 treatment effects through target trial emulation. Both groups designed SCMs to evaluate glucocorticoid treatment effects on 28-day mortality using real-world data from more than 2,000 hospitalized patients, benchmarked against published RECOVERY randomized controlled trial results. The best performing SCMs achieved bootstrap coverage rates exceeding 90% for two of three severity strata. Both human and AI models demonstrated equivalent clinical plausibility (n=3 expert reviewers) and similar statistical performance, though both struggled with critical disease severity. Ablation experiments comparing SCM-based approaches against traditional potential outcomes methods revealed SCMs achieved 76-98% coverage versus 1-37% for traditional methods. These results suggest that structural causal models can effectively bridge the interpretability gap in clinical AI by providing essential scaffolding for reliable causal inference and enabling meaningful human-AI collaboration while preserving methodological rigor essential for evidence-based medicine.</p>","PeriodicalId":34954,"journal":{"name":"Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing","volume":"31 ","pages":"309-323"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147310814","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Patch-level phenotype identification via weakly supervised neuron selection in sparse autoencoders for CLIP-derived pathology embeddings.","authors":"Keita Tamura, Yao-Zhong Zhang, Yohei Okubo, Seiya Imoto","doi":"10.1142/9789819824755_0051","DOIUrl":"10.1142/9789819824755_0051","url":null,"abstract":"<p><p>Computer-aided analysis of whole slide images (WSIs) has advanced rapidly with the emergence of multi-modal pathology foundation models. In this study, we propose a weakly supervised neuron selection approach to extract disentangled representations from CLIPderived pathology foundation models, leveraging the interpretability of sparse autoencoders. Specifically, neurons are ordered and selected using whole-slide level labels within a multiple instance learning (MIL) framework. We investigate the impact of different pre-trained image embeddings derived from general and pathology images and demonstrate that a selected single neuron can effectively enable patch-level phenotype identification. Experiments on the Camelyon16 and PANDA datasets demonstrate both the effectiveness and explainability of the proposed method, as well as its generalization ability for tumor patch identification.</p>","PeriodicalId":34954,"journal":{"name":"Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing","volume":"31 ","pages":"708-721"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147310826","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Leveraging Large Language Models to Derive Multiple Sclerosis Progression Assessments from Clinical Notes: A Feasibility Study.","authors":"Sy Hwang, Sunil Thomas, Heather Williams, Tom Hutchinson, Emily Schriver, Ashley Batugo, Amit Bar-Or, Vishakha Sharma, Frederik Buijs, Christopher Perrone, Danielle Mowery","doi":"10.1142/9789819824755_0023","DOIUrl":"10.1142/9789819824755_0023","url":null,"abstract":"<p><p>Ascertainment of multiple sclerosis (MS) progression is important for informing clinical care decisions and supporting biomedical research. However, the details to infer a patient's MS progression status are locked within clinical notes. In this feasibility study, we assessed the feasibility of developing and validating a large language model (LLM)-based EDSS and FS classifier for ascertaining MS progression from clinical notes.</p>","PeriodicalId":34954,"journal":{"name":"Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing","volume":"31 ","pages":"324-337"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147310830","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Systems Biology and Network Analysis: From Multi-omics Integration to Biological Mechanisms.","authors":"Gurkan Bebek, Onur Mutlu, Iman Hajirasouliha, Joshua Welch, Serguei Pakhomov","doi":"10.1142/9789819824755_0054","DOIUrl":"10.1142/9789819824755_0054","url":null,"abstract":"<p><p>The following sections are included: Session Introduction: Systems Biology and Network Analysis; Session Summary; Acknowledgments.</p>","PeriodicalId":34954,"journal":{"name":"Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing","volume":"31 ","pages":"752-754"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147310778","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Towards Automated Analysis of Gaze Behavior from Consumer VR Devices for Neurological Diagnosis.","authors":"Lio Schmitz, Markus Plack, Berkan Koyak, Muhammad Ehsan Ullah, Ahmad Aziz, Reinhard Klein, Zorah Lähner, Hannah Dröge","doi":"10.1142/9789819824755_0016","DOIUrl":"10.1142/9789819824755_0016","url":null,"abstract":"<p><p>Recent studies have demonstrated that eye tracking is a valuable tool in the detection, classification and staging of neurodegenerative diseases such as Parkinson's Disease (PD). However, traditional methods for capturing gaze data often rely on expensive and non-engaging clinical equipment such as video-oculography, limiting their accessibility and scalability. In this work, we investigate the feasibility of using eye tracking data collected via consumer-grade virtual reality (VR) headsets to support neurological diagnostics in a more accessible and user-friendly manner.This approach enables large-scale, low-cost, and remote assessments, which are particularly valuable in early detection and monitoring of neurodegenerative conditions. We show that relevant oculomotor features extracted from VR-based eye tracking can be used for predictive assessment. Despite the inherent noise and lower precision of consumer devices, careful preprocessing and robust feature engineering, including deep learning embeddings, mitigate these limitations. Our results demonstrate that both handcrafted and learned features from gaze behavior enable promising levels of classification performance. This research represents an important step towards scalable, automated, and accessible diagnostic tools for neurodegenerative diseases using ubiquitous VR technology.</p>","PeriodicalId":34954,"journal":{"name":"Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing","volume":"31 ","pages":"219-235"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147310761","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Applications of AI & ML in Biomanufacturing of Cell and Gene Therapies.","authors":"Eric Neumann, Karen Weisinger, Tom Londo","doi":"10.1142/9789819824755_0063","DOIUrl":"10.1142/9789819824755_0063","url":null,"abstract":"<p><p>This workshop highlights how AI/ML technologies are beginning to be applied to biomanufacturing and bioengineering of cell and gene therapies (CGT). AI/ML have demonstrated their utility in biocomputing and biomedical research applications, and are poised to become central to design, scaling, and optimization of bioengineering processes such as CAR-T cells, iPSC, and biomolecule production. Invited speakers from academia and industry will speak of their experience in leveraging these new intelligent technologies.</p>","PeriodicalId":34954,"journal":{"name":"Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing","volume":"31 ","pages":"855-858"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147310517","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Session Introduction: Biological molecular function: methods and benchmarks for finding function in biological dark matter.","authors":"Jason E McDermott, Yana Bromberg, Hannah Carter, Travis Wheeler","doi":"10.1142/9789819824755_0029","DOIUrl":"10.1142/9789819824755_0029","url":null,"abstract":"<p><p>The accurate determination of biological molecular function remains one of the most significant challenges in computational biology, with vast areas of biological \"dark matter\" persisting in microbiomes, viruses, and unexplored sequence space. To meet this challenge, we developed at PSB session to address the limitations of traditional sequence similarity-based functional annotation methods and explores how recent advances in AI/ML and high-throughput data generation are transforming the field. We highlight four innovative contributions presented in this session: a geometric framework using signed distance functions for modeling protein surfaces; a reinforcement learning-based approach for steering protein generative models to design functional sequences; an ensemble framework combining sequence, structural, and network features for subcellular localization prediction; and a scalable factorization method integrating gene-gene interaction data for analyzing high-dimensional genetic perturbation profiles. Together, these methodologies showcase the potential for computational and AI-driven tools to address the complex and multiscale nature of molecular function prediction, paving the way for new discoveries in understanding and engineering biological systems.</p>","PeriodicalId":34954,"journal":{"name":"Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing","volume":"31 ","pages":"417-424"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147309921","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Literature-driven extraction and computational prediction of causal statements linking genetic variants to biological processes, pathways and phenotypes.","authors":"Jici Jiang, Predrag Radivojac, Benjamin M Gyori","doi":"10.1142/9789819824755_0053","DOIUrl":"10.1142/9789819824755_0053","url":null,"abstract":"<p><p>Understanding the mechanistic basis of pathogenic genetic variants requires reconstructing the molecular pathways connecting the variant, via a chain of molecular intermediates, to a disease-causing biological process and phenotype. However, a literature-wide assembly of causal networks connecting variants, molecular pathways, biological processes and phenotypes has not been previously available. To create such a resource, we developed an automated pathway reconstruction approach building on the Integrated Network and Dynamical Reasoning Assembler (INDRA) system which extracts causal mechanistic statements (positive regulation, phosphorylation, complex formation, etc.) by combining structured databases and literature mining. We traversed INDRA statements extracted from publications to identify those describing a genetic variant resulting in a protein point mutation. We then reconstructed directed paths (consisting of one or more linked INDRA statements) connecting this variant to a term representing a biological process, phenotype or disease within the same publication. This resulted in a directed multigraph obtained from 25,862 paths for variants in 2,561 proteins. Each node in this graph corresponds to an ontology-grounded molecular or process term and each edge is explicitly linked to supporting literature evidence, enabling full auditability of inferred mechanisms. To leverage the assembled networks, we trained a classification model to predict likely downstream biological processes or specific disease associations for protein variants. As features to the model, we integrated molecular annotations (including protein sequence features, ClinVar pathogenicity labels, and UniProt domain mappings) in combination with representations from the ESM2 transformer-based protein language model. The performance achieved by this model shows promise for reconstructing causal mechanistic statements associated with function of genetic variants, a framing of the variant effect prediction task that goes significantly beyond simple assessment of pathogenicity. This integrative framework enables the mechanistic interpretation of known variants and prediction of functional relevance for variants lacking prior phenotypic annotation.</p>","PeriodicalId":34954,"journal":{"name":"Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing","volume":"31 ","pages":"738-751"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12952665/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147310862","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Asking the Right Questions: Benchmarking Large Language Models in the Development of Clinical Consultation Templates.","authors":"Liam G McCoy, David Wu, Sarita Khemani, Saloni Kumar Maharaj, Arth Pahwa, Leah Rosengaus, Lena Giang, Olivia Jee, Ethan Goh, Fateme Nateghi Haredasht, Kanav Chopra, David Jh Wu, Abass Conteh, Vishnu Ravi, Yingjie Weng, Kelvin Zhenghao Li, Daniel Shirvani, Jonathan H Chen","doi":"10.1142/9789819824755_0028","DOIUrl":"10.1142/9789819824755_0028","url":null,"abstract":"<p><p>This study evaluates the capacity of large language models (LLMs) to generate structured clinical consultation templates for electronic consultation. Using 145 expert-crafted templates developed and routinely used by Stanford's eConsult team, we assess frontier models-including o3, GPT-4o, Kimi K2, Claude 4 Sonnet, Llama 3 70B, and Gemini 2.5 Pro-for their ability to produce clinically coherent, concise, and prioritized clinical question schemas. Through a multi-agent pipeline combining prompt optimization, semantic autograding, and prioritization analysis, we show that while models like o3 achieve high comprehensiveness (up to 92.2%), they consistently generate excessively long templates and fail to correctly prioritize the most clinically important questions under length constraints. Performance varies across specialties, with significant degradation in narrative-driven fields such as psychiatry and pain medicine. Our findings demonstrate that LLMs can enhance structured clinical information exchange between physicians, while highlighting the need for more robust evaluation methods that capture a model's ability to prioritize clinically salient information within the time constraints of real-world physician communication. Limitations include reliance on Stanford-specific templates and concordancebased grading, which may not capture all clinically reasonable outputs.</p>","PeriodicalId":34954,"journal":{"name":"Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing","volume":"31 ","pages":"400-416"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13138846/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147310515","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Steering Protein Generative Models at Test-Time for Guided AAV2 Capsid Design.","authors":"Ben Viggiano, Wenhui Sophia Lu, Xiaowei Zhang, Luis S Mille-Fragoso, Xiaojing J Gao, Euan Ashley, Wing Hung Wong","doi":"10.1142/9789819824755_0031","DOIUrl":"10.1142/9789819824755_0031","url":null,"abstract":"<p><p>Recent advances in protein generative models have created new opportunities for protein engineering. However, a significant challenge remains in effectively steering these models to generate sequences with specific, desired functionalities, especially when these properties are defined by \"black-box\" or non-differentiable fitness functions. To address this, we present ProVADA+, a model-agnostic framework that guides pretrained generative models at testtime without costly retraining. Our approach introduces a reinforcement learning-based adaptive masking technique (MADA-DUCB) that significantly accelerates convergence. We demonstrate this framework on the challenging task of designing novel Adeno-Associated Virus 2 (AAV2) capsids. By coupling a ProteinMPNN generative prior with a fine-tuned AAV viability oracle, our method successfully navigates the rugged fitness landscape where unguided random mutagenesis is ineffective-with prior experiments showing as few as 0.3% of variants with six or more mutations are viable. In its final iterations, ProVADA generated a pool of novel candidates with a mean viral selection score of 2.72, consistently scoring highly viable variants while maintaining a diverse range of sequence similarity to the wildtype sequence. Our results show that ProVADA provides a powerful and efficient framework for accelerating the design of proteins with complex, user-defined properties.</p>","PeriodicalId":34954,"journal":{"name":"Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing","volume":"31 ","pages":"438-451"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12952671/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147310765","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}