Vladislav Leonov 1, Anna Kopeykina 1, Almina Polinova 1, Violetta Polinova 1, Elita Gerasimuk, Nikolay Pervushin 1, Nail Gizzatkulov 1, Ekaterina Mogilevskaya 1, Ekaterina Goryacheva 1, Oleg Demin 1
1 InSysBio UK Ltd (Edinburgh, UK)
Objectives: Quantitative Systems Pharmacology (QSP) model development requires extensive data for calibration and validation, yet sourcing these data from literature remains time-consuming and error-prone. Reported values are scattered across thousands of publications and are often incompatible due to inconsistent units, variable demographics, and heterogeneous statistical summaries. To address these challenges, we developed two freely accessible, manually curated web-based databases: CYTOCON DB (https://cytocon.insysbio.com) and FIVEDB (https://5db.insysbio.com). Both platforms are publicly available and aim to reduce the time spent on manual data extraction, standardize diverse literature data, and provide ready-to-use values for QSP modelers.
Methods: CYTOCON DB (Cell and cYTOkine CONcentrations Database) aggregates in vivo baseline concentrations of cytokines and immune cells from human samples. FIVEDB (cell process in vitro experiment database) collects in vitro experimental data describing cell death, differentiation, migration, proliferation, and secretion, and automatically translates them into kinetic parameters (e.g., kbase, Emax/Imax, EC50/IC50). Both databases are interactive web applications built on ASP.NET MVC with Microsoft SQL Server and Bootstrap. Shared core features include: (i) unit standardization—cytokine concentrations are converted to pM, cell counts to kcell/L, and time units to hours; (ii) normalization of reported statistics—all types of averages and dispersion measures are converted to mean and standard deviation; (iii) a two-step quality control workflow combining manual validation by annotators and reviewers with automated outlier detection; (iv) dynamic updates with over 1,000 new values added monthly from peer-reviewed sources; and (v) query-based retrieval with filters and export capabilities, including API access.
Results: The full CYTOCON DB contains 111601 concentration values for 790 molecules and 300 cell types across 548 disease types and 252 tissues, curated from 2901 publications. fIVE DB aggregates 41,945 kinetic parameters from 1,420 publications, covering secretion (40,302), cell death (833), migration (392), proliferation (306), differentiation (109), and cytokine degradation (3).
To illustrate the utility of CYTOCON DB, concentrations of IL-1α, IL-1β, IL-6, and TNF-α in serum from healthy controls and SLE patients were retrieved, visualized, and compared. TNF-α levels were notably higher in SLE (1.972±4.418 pM) than in healthy controls (0.928±3.305 pM), demonstrating the platform’s capacity to identify disease-specific differences in standardized units [1]. For fIVE DB, basal cytokine production constants for various immune cell types were retrieved and analyzed. The resulting cytokine profiles revealed high IL-8 secretion in stimulated monocytes and neutrophils, along with pronounced IL-6 and CCL2 secretion by monocytes, providing ready-to-use rate constants for QSP model calibration [2].
Conclusions: CYTOCON DB and FIVEDB together provide a comprehensive resource for the QSP modeling community. By combining in vivo baseline concentrations with in vitro derived kinetic parameters, these databases cover complementary aspects of model parameterization. Automated unit conversions, statistical normalization, rigorous quality control, and regular updates ensure data consistency and reliability. Free public access, combined with API support, promotes seamless integration into computational workflows and supports reproducible, transparent model development.
References:
[1] CYTOCON DB. InSysBio. https://cytocon.insysbio.com. 28.02.2026
[2] FIVEDB. InSysBio. https://5db.insysbio.com. 28.02.2026
Reference: PAGE 34 (2026) Abstr 12181 [www.page-meeting.org/?abstract=12181]
Poster: Software Demonstration