TrueGate Bio is an analytics and data science consultancy led by Kalif Shear, a cell and gene therapy scientist with 10+ years across public health research and FDA-regulated GMP development. The core of the practice is analytics: flow cytometry analysis (conventional and spectral), statistical analysis, and QC, stability, and method health data tools built in Python and R. Around it sits the regulated discipline that makes the numbers defensible: validation, method lifecycle governance, quality systems, and governed AI for GxP.
Analytics dashboards, statistical tools, and documents across flow cytometry analytics, data science, method lifecycle governance, quality and regulatory work, and governed AI. Demonstration and representative data only; no proprietary, employer, or client information.
The center of the practice: flow cytometry analytics for cell and gene therapy, conventional and spectral, from cross-site standardization and rare-event statistics to release gating and automated-gating validation.

Why precision for a 0.01% population has a floor set by counting statistics, how many events are needed for 20% and 35% CV, and how to write the LOQ into the method.

Daily bead QC across three sites, tested with Levey-Jennings charts and Westgard rules. Separates drift from variability and shows how a brighter channel moves a dim population across a fixed gate.

CAR-T immunophenotyping and release gating (%viable, %CD3, %CAR+, T-cell subsets, viable dose) plus spectral panel QC: full-spectrum signatures, a similarity matrix, and unmixing-risk flags before samples are run.

An independent QC overlay on gating across CAR-T phenotyping, PBMC immunophenotyping, CD19 residual impurities, and T-cell purity. Quantifies inter-analyst variability and flags disagreement over tolerance.

Validates automated gating against expert manual ground truth with R², bias, Bland-Altman limits, and event-level F1 per population, and flags where a human should still gate.
Worked analyses and data science tools built in Python and R: method comparison, equivalence testing, stability poolability, process capability, and significance testing. Each worked analysis uses synthetic data and walks from the question to the decision.

A comfortable Ppk over 40 batches hides a downward trend. The individuals chart, run rules, and recent-batch capability show the margin shrinking.

An ICH Q1E poolability test on three stability batches. Unequal slopes mean the shelf life comes from the worst batch, not from pooled data that overstate it.

Two transfers both pass a t-test; only one is equivalent. Two one-sided tests against a pre-set margin show why a small, noisy study cannot prove agreement.

Passing-Bablok and Bland-Altman show a proportional bias traced to a supercoiled plasmid standard, and what it would do to lots near a release limit.

Separates genuine off-target edits from sequencing noise with mismatch and CFD scoring, two-proportion tests, Wilson 95% CIs, and Benjamini-Hochberg FDR correction, cross-checked against SciPy.

Turns raw QC and stability data into decisions: cleaning, KPI cards, lot and attribute filters, out-of-spec and out-of-trend flags, and a shelf-life projection.
A six-exhibit packet following one analytical method from creation to the day someone has to prove its numbers can be trusted. Built the way I would build it for a decision meeting.

How the data behind a single laboratory measurement is governed: raw files, processing rules, settings, and the checks that decide whether a number can be trusted and defended years later.

One analytical method traced from development through qualification, transfer, routine QC, investigation, CAPA, and revalidation, with the entities, state transitions, and relationship rules at each step.

A decision-meeting deck on a method that passed transfer and still failed in routine use: the risk assessment that found the structural cause, the controls that fixed it, and the evidence the fix held.

A one-page post-transfer monitoring dashboard: signals and thresholds agreed in advance, recovery against a pre-set criterion, and classified root causes with escalation rules.

A two-page memo to analytical leadership and the quality council requesting a transfer-readiness gate and a post-transfer monitoring window, with ownership and source-of-truth rules.
The supporting register behind the packet: every failure mode scored, then re-scored after the controlled change, with live formulas so the change in risk is shown rather than asserted.
Validation, compliance, training, and regulatory strategy documents: the structures I use when an answer has to survive an auditor or an agency.

The tracker structure for proving a set of external requirements satisfied, item by item, to someone who will audit the answer: two-way traceability, gaps as owned records, and independent closure evidence.

A 15-slide training deck on electronic records and signatures for flow cytometry labs: audit trails, access control, data integrity, and what an inspector actually checks.

Twelve GxP controls set against their equivalents in a regulatory inquiry function: same failure mode, same evidence, same reason the control exists.

A risk-stratified framework for finding which CMC gaps change the pathway, timeline, or risk of an IND filing, and which can proceed in parallel.

Deciding whether a Pre-IND meeting is the right move, and how to make it count: readiness questions, question framing, and briefing package structure.

A template for classification, pathway, and milestones for a novel therapeutic, built to document the reasoning FDA will test, not just the conclusion.
Governed, human-in-the-loop AI tools for the highest-burden GMP documentation tasks. Each drafts fast, checks itself, and flags every point that needs a qualified human.

Turns process notes into a review-ready SOP draft, generates a QC checklist, and flags every point that needs a human decision. Nothing is approved until a qualified reviewer signs off.

Turns an event description into a structured deviation and CAPA: impact assessment, investigation, root cause, action plan, and effectiveness check. Reportability and disposition stay human decisions.

Turns validation study data into a structured report with results against acceptance criteria (accuracy, precision, linearity, LOD/LOQ, robustness), aligned to ICH Q2(R2). The validated conclusion stays human.

A controlled-document writing sample: a flow cytometry viability SOP structured for GMP use, step by step and audit-minded, with data-integrity requirements built in.
Run a Passing-Bablok method comparison, size a study, check what a spot-check proves, or estimate shelf life, right in your browser.
Open the tools →How real problems were solved: confounded field data, automated classification in production, a validated system migration, and building QC from nothing.
Read the case studies →Short notes on method transfer, method comparison statistics, flow cytometry data integrity, and governing AI in change control.
Read the insights →Analytics and data science first, for biotech, pharma, and IVD teams, grounded in hands-on GMP QC, analytical development, and regulatory work.
Flow cytometry analysis (conventional and spectral), statistical analysis, method health and QC or stability data tools, built in Python, R, JMP, and Prism.
Method validation and transfer, computerized system validation, 21 CFR Part 11 gap assessments, traceability matrices, and CAPA with effectiveness criteria set in advance.
SOPs, validation and investigation reports, decision memos, CMC content, and IND-readiness frameworks, written to survive review by an auditor or an agency.
Human-in-the-loop AI workflows for documentation, plus evaluation criteria and acceptance standards for AI output. The human review step is designed in, not bolted on.
I founded TrueGate Bio as an analytics and data science consultancy because the hardest part of cell and gene therapy data is making the numbers defensible. I specify what correct looks like, then prove a system meets it. Ten-plus years across public health research at the CDC and FDA-regulated cell and gene therapy at Cellphire, Fate Therapeutics, and Kite Pharma/Gilead: method and computerized system validation, method lifecycle management, quality systems, root cause investigation, supplier oversight, and regulatory documentation under ICH and FDA expectations.
I apply the same discipline to AI systems: evaluation criteria and rubric authorship, reward and scoring function design, adversarial scenario design, ground-truth construction, and model output assessment with written rationale. Defining acceptance criteria before you trust an output is a regulated practice, not an instinct, which is why the governed AI tools here keep a qualified human in the loop.
I work across quality systems, validation, data integrity, laboratory informatics, and analytical development, and I build the software when the right tool doesn't exist. Download my CV →
Available for analytics and data science engagements, flow cytometry analysis, validation, regulatory writing, and governed AI for GxP teams. Open to the right full-time role as well.