A detailed look at the architecture, mathematical models, intelligent gating algorithms, and enterprise-grade cloud security that power our browser-based cell cycle analysis platform.
Unlike legacy cytometry software that ties you to specific operating systems, hardware dongles, or expensive yearly licenses, CellCycle runs a full scientific Python stack directly inside your browser using Pyodide - a WebAssembly port of CPython. This paradigm shift means industry-standard data science libraries like NumPy, SciPy, and our innovative analysis engine execute at near-native speed on your local CPU, completely eliminating server dependency and latency.
The entire high-throughput analytical pipeline is optimized for performance. From initial FCS 3.1 binary parsing, through complex multi-dimensional histogram construction, robust peak detection, and non-linear curve fitting, to the final statistical rendering - the process completes in under 2 seconds on a standard consumer laptop. There is no software to install, no complex IT deployment approvals required, and no risk of version mismatch across your research team.
Your raw .fcs files never leave your machine, ensuring immediate compliance with HIPAA and GDPR regulations. All computationally intensive analysis happens entirely within your browser's sandboxed memory. When you close the tab, the data is completely purged from memory - leaving absolutely zero traces on any external server.
Flow cytometry analysis is no longer restricted to Windows workstations. CellCycle works identically and flawlessly on Windows, macOS, Linux, and even ChromeOS. Any modern web browser with WebAssembly support (Chrome, Firefox, Edge, Safari) instantly becomes a powerful analytical workstation.
We leverage the same computational backbone trusted by leading data scientists. Powered by NumPy for high-performance array operations, SciPy for advanced signal processing and non-linear optimization, and Plotly.js for hardware-accelerated, highly interactive 2D and 3D data visualizations.
Designed for large-scale cohort experiments. Load dozens of FCS files simultaneously into your workspace. The WebAssembly engine processes them sequentially, applying consistent auto-gating parameters across the entire dataset, and automatically generating comprehensive, combined multi-page PDF and CSV reports.
Accurate cell cycle quantification is a mathematically complex challenge requiring the precise separation of overlapping G1, S, and G2/M phase distributions from a single, often noisy, DNA fluorescence histogram. To achieve clinical-grade precision, CellCycle implements a suite of advanced deconvolution algorithms. Each model utilizes rigorous Non-Linear Least Squares (NLLS) optimization, powered by SciPy's curve_fit with bounded constraints, to autonomously converge on the absolute best mathematical fit for your specific biological sample.
Our primary, highly versatile deconvolution model. It utilizes a broadened, second-degree polynomial to accurately represent the S-phase distribution between the G1 and G2 Gaussian peaks. The DJF model is exceptionally robust at adapting to complex, non-linear variations in DNA synthesis rates across diverse mammalian cell lines and experimental treatment conditions.
An alternative, elegant mathematical approach that employs analytical approximations to seamlessly bridge the Gaussian tails of the G1 and G2/M populations. The Watson model is renowned for offering highly stable, rapidly converging fits, making it particularly well-suited for analyzing clean, high-resolution flow cytometry data with distinct phases.
Simple trapezoidal debris subtraction is often inadequate. All of our core models are fundamentally augmented with a Complementary Error Function (ERFC) plateau. This advanced mathematical component accurately describes the "sliced background" - the flat debris shelf that extends to the left of the G1 peak - effectively preventing the critical overestimation of the S-phase fraction.
Unlike conventional cell cycle analysis software, every fit generated by CellCycle undergoes explicit quantitative validation. The RCS metric is computed using data-driven Poisson variance exclusively within a defined Region of Interest (ROI) mask. Restricting the analysis to biologically relevant regions minimizes instability introduced by low-information tail noise, allowing RCS values between 1.0 and 5.0 to reflect strong agreement between observed and fitted distributions.
Manual polygon gating introduces significant intra- and inter-operator variability, inherently compromising the reproducibility of large-scale, multi-center studies. CellCycle's auto-gating engine replaces subjective human estimation with a deterministic, multi-layered statistical approach. It autonomously isolates the true singlet cell population, rigorously filtering out subcellular debris, multi-cell aggregates, and detector saturation artifacts - guaranteeing 100% reproducibility across your entire dataset.
Rather than applying static gating templates that frequently fail when experimental conditions shift, our algorithm dynamically analyzes the unique FSC/SSC density topology of every individual sample. It calculates precise statistical boundaries that adapt intelligently to varying instrument noise floors, subtle voltage shifts, and inherent biological sample variability, all without requiring manual user intervention.
The process begins by computing a high-resolution 2D histogram of the Forward Scatter (FSC) versus Side Scatter (SSC) parameters. Advanced Gaussian smoothing matrices are then applied to clearly reveal the dense core of the biological population cluster, effectively suppressing isolated background noise events and harsh saturation-wall artifacts at the instrument's detection limits.
Instead of drawing arbitrary boxes, the engine robustly estimates the covariance matrix of the core population. It then calculates every single event's multidimensional Mahalanobis distance from the population centroid. Events exceeding a strict statistical threshold (χ² = 25) are objectively classified as outliers and excluded, forming a perfectly contoured elliptical gate.
Immediately after gating, the engine scans the resulting DNA fluorescence histogram to autonomously locate both the G1 and G2/M peaks using highly tuned scipy.signal.find_peaks algorithms. This logic is specifically designed to ignore zero-channel debris spikes and strictly search within the biologically valid 1.6×–2.4× G1 positional window to locate the corresponding G2 peak, even in aneuploid samples.
While the engine is fully autonomous, the scientist always remains in complete control. Users can instantly override or refine the automatically generated boundaries using interactive, reactive UI sliders (Gate Width, G2 Reach, Debris Cut) or by drawing a manual polygon. Because the computation runs locally in WebAssembly, millions of events are re-gated and re-visualized in milliseconds.
In cytometry, black-box algorithms are unacceptable. CellCycle is built on the philosophy of absolute computational transparency. Every step of the analysis, from initial debris segmentation to final S-phase modeling, is fully deterministic, mathematically exposed, and strictly verifiable.
Every fitted component, statistical assumption, and covariance matrix remains exposed. You see exactly why a population was classified as G2/M.
We don't just provide the output; we quantify its validity. Reduced Chi-Square (RCS) metrics and residual plots are continuously reported for every sample.
Automated segmentation is a starting point, not a mandate. Manual override is always available, and all automated gating boundaries remain strictly user-adjustable.
The pipeline relies entirely on published, peer-reviewed mathematical frameworks (Dean-Jett-Fox, Theil-Sen regression) - zero reliance on opaque neural networks or unpredictable AI inference.
Data security and patient privacy are paramount in modern biomedical research. CellCycle's core analytical engine is built upon a strict Zero-Trust, privacy-first architecture: absolutely all raw FCS data computation occurs locally within your browser's sandboxed environment. However, for clinical and enterprise users requiring permanent data archiving and strict regulatory compliance, we provide an optional, seamlessly integrated cloud tier powered by Google Firebase.
This cloud infrastructure is the same robust backbone trusted by top-tier financial and healthcare institutions globally. Google Cloud guarantees enterprise-grade encryption both at rest (AES-256) and in transit (TLS 1.3), maintains rigorous SOC 2/3 compliance, and delivers a 99.95% uptime SLA. When you choose to archive an analysis, your generated PDF and CSV reports inherit the exact same formidable security posture that protects Google's own internal products.
Designed to meet FDA 21 CFR Part 11 requirements. Every significant analytical action - ranging from algorithm model selection, manual gating overrides, peak coordinate adjustments, to final file exports - is securely tracked and logged in Firebase Firestore with precise timestamps and user identities. These comprehensive audit records are immutable; they cannot be altered or deleted post-creation, ensuring full, transparent regulatory traceability for every experiment.
Finalized reports can be cryptographically "Signed & Locked" using advanced Firebase Authentication protocols. This action irrevocably binds the verified identity of the supervising analyst to the specific mathematical state of the analysis. Once a record is signed and locked, it becomes entirely tamper-proof, preventing any subsequent modifications and ensuring data integrity for publication or regulatory submission.
Say goodbye to scattered local hard drives and lost flash drives. Your generated PDF analytical reports and comprehensive CSV statistical datasets are securely stored in private Google Cloud Storage buckets. Access your historical data instantly from any device worldwide, easily share authenticated read-only links with global collaborators, and maintain a centralized, permanent, and backed-up record of all your laboratory's flow cytometry analyses.
Platform access and user identity authentication are seamlessly handled by Firebase Auth, fully supporting robust email/password configurations as well as Google Single Sign-On (SSO) for enterprise environments. Our meticulously crafted Firestore Security Rules act as an impenetrable firewall, enforcing strict per-user data isolation at the database level - mathematically guaranteeing that no user can ever query or access another researcher's confidential reports or audit logs.
No registration required. Drop a .fcs file and see the full analysis pipeline in action - directly in your browser.
Try CellCycle Free