Computational and structural biotechnology journal最新文献

{"title":"QuRad: Radiomics Feature Extraction for QuPath.","authors":"Janan Arslan, Tamara Künzle, Yeşim Güner, Agustí Alentorn","doi":"10.34133/csbj.0238","DOIUrl":"https://doi.org/10.34133/csbj.0238","url":null,"abstract":"<p><p>Radiomic features provide a quantitative description of tissue organization in histopathology images. QuPath, a widely used open-source platform for whole-slide image analysis, offers morphology, intensity, and Haralick texture measurements but no standardized radiomic feature set such as that of PyRadiomics. Researchers must therefore export data to external Python pipelines, interrupting their QuPath workflow. QuRad extracts radiomic features from cell detections and annotation regions directly inside QuPath. The extension and equivalent Groovy script implement 119 features: 103 two-dimensional features enabled by default and 16 optional legacy shape descriptors. We compared QuRad 0.4 with PyRadiomics 3.0.1 on identical grayscale images and pixel masks. All 103 default features met a 10^-9 absolute-or-relative agreement tolerance on a breast-cancer benchmark tile (410 cells). On 20 melanoma tiles (8,644 nuclei), 97 of 103 features met this tolerance and 6 polygon-derived shape features differed from PyRadiomics' raster-derived mesh (minimum concordance correlation coefficient 0.942). The 22 synthetic regions accepted by PyRadiomics reproduced the expected intensity and texture values. A single-threaded server benchmark reached about 1,900 nuclei per second, and processing the 184,242 cells of a whole-slide image took 2.7 min on a laptop. A random forest distinguished tumor cells from lymphocytes in melanoma (pooled macro-F1 0.853) and classified 197 tissue regions of 6 breast-cancer slides into 3 compartments (0.696). QuRad exports the features with object coordinates and extraction settings for use in QuPath or downstream analyses.</p>","PeriodicalId":10715,"journal":{"name":"Computational and structural biotechnology journal","volume":"35 1","pages":"0238"},"PeriodicalIF":4.8,"publicationDate":"2026-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13678076/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149223516","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The Current Status and Evaluation of Artificial Intelligence in Bladder Cancer: Clinical Applications and Research.","authors":"Zhuo Yin, Joshua W K Ho","doi":"10.34133/csbj.0197","DOIUrl":"https://doi.org/10.34133/csbj.0197","url":null,"abstract":"<p><p>Bladder cancer ranks as the 10th most common malignancy worldwide. Advances in technology are continuously reshaping its diagnosis, treatment selection, and prognosis assessment. In recent years, artificial intelligence (AI) has demonstrated marked potential in transforming bladder cancer care. By integrating and analyzing multi-source data, AI enhances the precision and efficiency of disease management. This review comprehensively examines the current applications of AI across various domains, including radiographic imaging analysis, pathological diagnosis, cystoscopic image processing, molecular subtyping, drug development, organoid research, prognostic modeling, and medical device innovation. It highlights the advantages of AI over conventional methods, outlines model development strategies, and summarizes commonly employed algorithms. Furthermore, we discuss the key challenges in translating AI tools into real-world clinical practice. This narrative review aims to inform the future development of AI-driven digital health systems, ultimately contributing to the realization of precision medicine and personalized management for bladder cancer patients.</p>","PeriodicalId":10715,"journal":{"name":"Computational and structural biotechnology journal","volume":"35 1","pages":"0197"},"PeriodicalIF":4.8,"publicationDate":"2026-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13678086/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149223498","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Computational Design of a TCR-like Human V_H Domain Antibody That Specifically Recognizes a WT1-Derived Peptide Presented by Class I Major Histocompatibility Complex.","authors":"Tae-Sung Oh, Seong-Woo Kim, SangPhil Ahn, Seonghyuk Suh, JooYeon Lee, YeongNam Bae, Hyejin Lee, Yong-Sung Kim, Bo-Seong Jeong, Byung-Ha Oh","doi":"10.34133/csbj.0236","DOIUrl":"10.34133/csbj.0236","url":null,"abstract":"<p><p>T cell receptor (TCR)-like antibodies that recognize a peptide-major histocompatibility complex (pMHC) can enable precise targeting of cancer cells, but developing single-domain binders with high specificity and affinity is challenging. Here, we report the computational design and experimental validation of a TCR-like human heavy chain variable (V_H) domain antibody that specifically recognizes the Wilms' tumor 1 protein (WT1)-derived peptide RMFPNAPYL presented by HLA-A*02:01/β2-microglobulin (RMF/HLA-A*02). Starting from the crystal structure of a Fab antibody bound to RMF/HLA-A*02, we repurposed the soluble trastuzumab-derived V_H scaffold to bind RMF/HLA-A*02. The resulting V_H domain, αRMF_V_H, spans all 9 peptide residues and showed no binding to a control pMHC with a different peptide. Fusion of 1 or 2 αRMF_V_H molecules to human serum albumin domain III yielded monovalent and bivalent constructs with good biophysical properties. The monovalent construct bound selectively to RMF/HLA-A*02 with 81 nM affinity, and when expressed as part of a chimeric antigen receptor on T cells, it functioned as an antigen-recognition domain to selectively activate T cells in the presence of RMF/HLA-A*02-positive cells. The bivalent construct exhibited a dramatically increased apparent binding affinity of 0.37 nM. When both constructs were engineered into the bispecific T cell engager (TCE) format containing an anti-CD3 Fv, the bivalent TCE induced antigen-dependent cytotoxicity much more effectively than the monovalent TCE. These results demonstrate a viable strategy to develop high-specificity, peptide-focused TCR-like V_H domain antibodies that can be readily extended into potent bivalent TCEs for pMHC-targeted therapeutics.</p>","PeriodicalId":10715,"journal":{"name":"Computational and structural biotechnology journal","volume":"35 1","pages":"0236"},"PeriodicalIF":4.8,"publicationDate":"2026-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13669508/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149215203","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Lipid-Metal-Organic Framework Colloidosomes for Multimodal Encapsulation and Pollutant Capture.","authors":"Jeshua Podliska, Rahul Dev Jana, Raheleh Ravanfar","doi":"10.34133/csbj.0235","DOIUrl":"10.34133/csbj.0235","url":null,"abstract":"<p><p>The development of stable colloidosomes capable of accommodating chemically distinct molecular species while retaining accessible porosity remains an important challenge for multifunctional encapsulation and pollutant-capture applications. Here, we develop a hybrid lipid-metal-organic framework (lipid-MOF) colloidosome assembled through an interfacial emulsification strategy that integrates the structural rigidity of zeolitic imidazolate framework-8 particles with lipid-mediated membrane stabilization. During assembly, zeolitic imidazolate framework-8 particles accumulate at the oil-water interface to form a shell, producing hollow micron-sized spherical colloidosomes. The resulting colloidosomes exhibit excellent colloidal stability in aqueous media for more than 30 d with a zeta potential of approximately -50 mV. Nitrogen adsorption measurements reveal a surface area of 45 m² g^-1 and an average pore width of 4 nm for the dried composite following vacuum degassing at 100 °C for 24 h. Fluorescence imaging shows that hydrophobic Nile Red preferentially partitions into the colloidosomal membrane, whereas hydrophilic fluorescein isothiocyanate (t) localizes predominantly within the aqueous interior. The lipid-MOF colloidosomes enable encapsulation of hydrophilic methylene blue molecules with loading efficiencies approaching 90%. Furthermore, the colloidosomes demonstrate rapid removal of model pollutants from water, achieving >90% removal of methylene blue and metal ions without stirring. Together, these results introduce lipid-MOF colloidosomes as a new class of hybrid platforms that unify structural stability, multicompartmental encapsulation, and efficient adsorption behavior, opening pathways toward sustainable platforms for drug delivery application and environmental bioremediation.</p>","PeriodicalId":10715,"journal":{"name":"Computational and structural biotechnology journal","volume":"35 3","pages":"0235"},"PeriodicalIF":4.8,"publicationDate":"2026-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13662271/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149178815","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exact Formula of the Total Quasi-Steady State Approximation in Competitive Target-Mediated Drug Disposition.","authors":"Taehong Kim, Jaeyun Cha, Hyeseon Jeon, Hwi-Yeol Yun, Dongju Lim, Jae Kyoung Kim","doi":"10.34133/csbj.0223","DOIUrl":"10.34133/csbj.0223","url":null,"abstract":"<p><p>Competitive target-mediated drug disposition (competitive TMDD) arises when 2 drugs compete for the same target receptor. The full competitive TMDD model characterizes these dynamics; yet, its complexity motivated the use of a reduced model, which is invalid under high receptor concentrations. While this problem can be resolved by using the total quasi-steady state approximation, which remains valid for broad parameter regimes, the exact formula of the total quasi-steady state approximation-based reduced model-competitive qTMDD-remained unknown for 15 years, as it requires solving a cubic equation and identifying a biologically meaningful solution. Consequently, researchers have relied on computationally expensive numerical approximations, limiting practical applicability. Here, we derive-for the first time-a real-valued exact formula for competitive qTMDD by leveraging analytic properties of cubic equations and geometric characteristics of their roots in the complex plane. This formulation improved computational speed by approximately 18-fold compared to numerical approximation methods, enabling competitive qTMDD-based Bayesian inference that was previously impractical. Applied to clinical trial data for anakinra and rhIL-7-hyFc, competitive qTMDD estimated pharmacological parameters comparable to those from the full competitive TMDD model while requiring only approximately 28% of the computation time. Importantly, it maintains robust accuracy even at low sampling density, whereas previous reduced models exhibit bias. By ensuring precise biological interpretation of drug systems even in complex real-world scenarios, the exact formula of competitive qTMDD has the potential to streamline the drug development and clinical testing process. Our exact formula also consists entirely of real-valued terms, allowing seamless integration into existing pharmacometrics software.</p>","PeriodicalId":10715,"journal":{"name":"Computational and structural biotechnology journal","volume":"35 1","pages":"0223"},"PeriodicalIF":4.8,"publicationDate":"2026-10-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13646460/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149135281","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Unbiased Long-Read Whole-Genome Sequencing Enables High-Resolution Mapping of Transgene Concatenation and Off-target Genomic Disruption in a Mouse Model.","authors":"Maanik Mehta, Kashif Ahmed, Rowaida Hussein, Erika Tavares, Zorana Berberovic, Rabiat Adele, Abigail D'Souza, Bin Gu, Michael D Wilson, Evgueni Ivakine, Philippe P Monnier, Elise Héon, Ajoy Vincent","doi":"10.34133/csbj.0234","DOIUrl":"10.34133/csbj.0234","url":null,"abstract":"<p><p>Transgenic mouse models are indispensable for dissecting disease mechanisms, yet their interpretability is frequently compromised by cryptic genomic alterations introduced during transgenesis. Thus, robust quality-control strategies are needed to elucidate integration architecture and evaluate model performance when such unintended events occur. Here, we applied unbiased whole-genome long-read sequencing using the PacBio Revio to investigate a mouse model exhibiting unexpected transgene silencing, originally designed to recapitulate autosomal-dominant hereditary macular dystrophy driven by the up-regulation of a <i>ZZEF1-ALOX15</i> fusion gene. Long-read sequencing analysis revealed a ≥29-kb head-to-tail concatemer containing at least 3 transgene-derived copies. Reconstruction of transgene-genome junctions revealed unintended off-target integration of the concatemer into the calcium-sensing receptor gene (<i>Casr</i>), along with exogenous <i>Escherichia coli</i> DNA, that together defined final transgene architecture. Aggregate 5-methylcytosine profiling identified hypermethylation of the transgene promoter across the concatemer, and additional phenotyping suggested disruption of endogenous CaSR function resulting from the rearrangement. Our workflow enabled direct detection of transgene concatenation and off-target mapping. These findings highlight the power of whole-genome long-read sequencing in uncovering hidden genomic complexity, resolving aberrant phenotypes, and enhancing the reliability of in vivo disease modeling, supporting its incorporation as a potentially scalable quality-control tool for genetically engineered animal models.</p>","PeriodicalId":10715,"journal":{"name":"Computational and structural biotechnology journal","volume":"35 1","pages":"0234"},"PeriodicalIF":4.8,"publicationDate":"2026-10-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13646464/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149135258","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Intraoperative Trajectory Realignment Algorithm for Minimizing Screw Misplacement in Cervical Spinal Fusion.","authors":"Pasit Komonsuwan, Jackrit Suthakorn, Shen Treratanakulchai, Sorayouth Chumnanvej","doi":"10.34133/csbj.0222","DOIUrl":"10.34133/csbj.0222","url":null,"abstract":"<p><p>Spinal fusion is a surgical procedure designed to stabilize the spine by connecting vertebrae with screws and rods, where placement accuracy is critical, as even small deviations can cause severe neurovascular complications. The accuracy of traditional preoperative image-based navigation is limited by intraoperative alignment shifts, particularly in the highly mobile cervical spine. This study presents an intraoperative trajectory realignment algorithm that reconciles preoperative computed tomography data with simulated and phantom-derived intraoperative 2-dimensional (2D) projections by integrating forward kinematics, digitally reconstructed radiographs, and optimization-based search strategies. The algorithm was evaluated through computational simulations and physical experiments using a modular 3D-printed cervical spine phantom, with accuracy evaluated through clinical grading and assessment metrics, including entry/end point displacement, angular deviation, and minimum inter-trajectory distance. Results demonstrated high accuracy, with all screws graded as perfectly placed and deviations consistently below clinically relevant thresholds. While these findings confirm the feasibility of this approach under controlled experimental conditions, future validation using cadaveric models and real intraoperative fluoroscopic imaging remains necessary prior to clinical integration.</p>","PeriodicalId":10715,"journal":{"name":"Computational and structural biotechnology journal","volume":"35 2","pages":"0222"},"PeriodicalIF":4.8,"publicationDate":"2026-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13643026/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149126882","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Design and Heterologous Production of Bagworm Silk-Inspired Proteins in <i>Aspergillus oryzae</i>.","authors":"Adiphol Dilokpimol, Ulrike Abendroth, Mari Nyyssönen, Anna S Borisova, Anni Seisto, Jenni Muli, Dominik Mojzita, Fiona Småros, Hans Mattila, Antti P Aalto, Mari Valkonen, Pezhman Mohammadi","doi":"10.34133/csbj.0233","DOIUrl":"https://doi.org/10.34133/csbj.0233","url":null,"abstract":"<p><p>Silk-inspired proteins are attractive biomacromolecules owing to their structural diversity, sustainability, and tunable physicochemical properties. Among natural silks, the bagworm silk produced by <i>Eumeta variegata</i> has attracted increasing attention because of its unique sequence architecture and remarkable reported mechanical performance. However, the recombinant production of bagworm silk-inspired proteins remains largely unexplored. In this study, we designed and heterologously expressed 2 bagworm silk-inspired protein variants in <i>Aspergillus oryzae</i> using a synthetic expression system. The variants comprised a representative bagworm silk with an alanine- and glycine-rich repeat motif flanked by carbohydrate-binding modules at the N- and C-termini. Expression screening in engineered <i>A. oryzae</i> strains demonstrated stable accumulation of C-tag-reactive species corresponding to both variants in a protease-deficient host, whereas comparable accumulation was not detected in a carbon-catabolite-repressor-deficient strain under the tested conditions. The selected variant was further produced in a 2-l bioreactor, demonstrating the feasibility of scale-up and establishing <i>A. oryzae</i> as a promising platform for bagworm silk-inspired protein production. The recovered H-eBGW-containing protein fractions were readily processed into thin films and exhibited temperature-dependent changes in mechanical properties and surface wettability, demonstrating the processability and curing-responsive behavior of the recovered material. Collectively, our results demonstrate the feasibility of producing structurally complex bagworm silk-inspired proteins in filamentous fungi and provide insights into the design, expression, and processing of recombinant silk-inspired proteins for future bio-based applications.</p>","PeriodicalId":10715,"journal":{"name":"Computational and structural biotechnology journal","volume":"35 1","pages":"0233"},"PeriodicalIF":4.8,"publicationDate":"2026-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13624362/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149105321","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Boolean Seven-Layer Platelet Decision Model Models Consistently the Logic and Semiquantitatively the Dynamics of Platelet Activation and Inhibition.","authors":"Özge Osmanoglu, Johannes Balkenhol, Juan Prada, Martin Kaltdorf, Maral Afshar, Sebastian Mammen, Katrin G Heinze, Stepan Gambaryan, Harald Schulze, Thomas Dandekar","doi":"10.34133/csbj.0228","DOIUrl":"https://doi.org/10.34133/csbj.0228","url":null,"abstract":"<p><p>A platelet decision model integrates available experimental data to consistently model the Boolean logic in line with all available pathophysiological experimental platelet data. Due to insufficient kinetic data on involved protein networks, we semiquantitatively describe the dynamics of platelet activation and inhibition, i.e., the order of events and the relative strength of receptor activation or protein phosphorylation. The Boolean network model (76 nodes, 115 interactions) incorporates 7 layers of platelet signaling and regulatory feedback loops: (a) platelet activation stimuli resulting from vascular injury, (b) the platelet receptor layer (GPVI, GPIb-IX-V, PAR1/4, P2Y₁/P2Y₁₂, P2X1, TP, 5-HT2A, and IP) that receives signals, (c) central kinases (Src, Fyn, Lyn, PKC, PKA, and PKG) that converge diverse signals, (d) the translation of signaling into changes in platelet ion homeostasis, (e) alterations in platelet shape/adhesion, (f) the activation of other platelets, and (g) the formation of a platelet plug. Dynamic simulations based on curated omics, genetics, and functional data from experiments demonstrate network robustness by multiple and redundant positive feedback mechanisms. These allow rapid and irreversible aggregation and make the platelets resilient to even genetic knockouts (e.g., Trpm7). Signaling in chronic inflammatory conditions is modeled, such as in atherosclerosis, inflammatory arthritis, and aging, as well as inherited platelet disorders. We describe how acute decision responses can become maladaptive leading to chronic inflammatory decisions: This implies a risk of thrombo-inflammatory and cardiovascular diseases.</p>","PeriodicalId":10715,"journal":{"name":"Computational and structural biotechnology journal","volume":"35 1","pages":"0228"},"PeriodicalIF":4.8,"publicationDate":"2026-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13624728/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149105346","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Machine Learning in Stem Cell Research: From Biological Data to Clinical Translation.","authors":"Anna Nicolaou, Yan Hong, Delong Zhou, Geovanni Geukgeuzian, Vasisht Yegneshwaran, Nehal Kamal Ali, Diego Fraidenraich, Xueqing Huang","doi":"10.34133/csbj.0212","DOIUrl":"https://doi.org/10.34133/csbj.0212","url":null,"abstract":"<p><p>Stem cell (SC) research plays a central role in disease modeling and regenerative medicine, yet its clinical translation remains challenged by biological heterogeneity, incomplete cellular maturation, stringent quality-control requirements, and increasingly complex biological and clinical data. Machine learning (ML) has emerged as a powerful computational approach for extracting biologically meaningful information from these data, enabling quantitative characterization of SC behavior and predictive modeling across diverse experimental and translational applications. This review presents a unified framework for ML in SC research by organizing the literature according to SC data sources and representations, ML methodologies and model families, and biological and clinical application domains. We review ML applications spanning SC reprogramming, differentiation, postdifferentiation maturation, quality control, disease modeling, therapeutic development, and clinical decision support, highlighting how diverse experimental data, including molecular and cellular profiling, biomedical imaging, biochemical profiling, functional measurements, and clinical data, enable biological interpretation, predictive modeling, and translational decision-making. We further discuss benchmark dataset selection, validation strategies, transformer-based foundation models, and emerging multimodal learning approaches and propose benchmark dataset selection criteria emphasizing biological diversity, experimental rigor, metadata completeness, reproducibility, and public accessibility. Collectively, this review provides a comprehensive framework for understanding current ML applications in SC research and outlines future directions toward robust, interpretable, and clinically translatable computational frameworks.</p>","PeriodicalId":10715,"journal":{"name":"Computational and structural biotechnology journal","volume":"35 1","pages":"0212"},"PeriodicalIF":4.8,"publicationDate":"2026-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13624364/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149105316","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}