Data without a clear conclusion
The data exist, but the analysis or biological interpretation is stalled.
Biological & biotechnology R&D × data science
Turning complex science into clear decisions.
Bring me in when an analysis is stalled, an experiment needs an independent review, or complex results must become a clear manuscript, figure set, or R&D decision. I combine experimental biology, genomics, statistics, reproducible R/Python, and fit-for-purpose machine learning.
15+years in scientific research
Co-inventorpublished international patent application
Scienceto signal
Between the datasets
Connecting the evidence behind the decision.
Most R&D teams have more data than time. Analyses get postponed, designs get locked in under deadline, and experiments can end up addressing something slightly different from what originally mattered.
The evidence behind an important decision rarely lives in one dataset. Sequencing data may sit with one team, phenotyping or imaging with another, and molecular assays, analytical measurements, or process data somewhere else again.
My approach is simple: start with the biological question, make uncertainty visible, and build only the analysis the decision requires.
Across Puregene AG, the Salk Institute, and Yale University, I have worked at the intersection of genome editing, genomics, molecular biology, imaging, statistics, and machine learning—from applied R&D to fundamental questions involving DNA replication and chromatin.
Bridging those gaps is the work.
Public examples use only open, synthetic, or independently created materials. No confidential or proprietary information is included.
When to bring me in
Concrete problems. Focused support.
Before an expensive experiment
The design, controls, endpoints, or analysis plan need an independent review.
Results need a scientific story
A manuscript, report, or figure set needs a clearer evidence-based narrative.
Evidence is split across teams
Genomics, phenotyping, imaging, and molecular results do not yet support one decision.
Consulting process
Clear scope. Useful handoff.
Initial conversation
Clarify the scientific question, priorities, and desired outcome.
Defined scope
Agree on deliverables, timing, and boundaries.
Focused work
Complete the analysis, review, figures, writing, or experimental guidance.
Useful handoff
Deliver documented results, reusable files, and clear next steps.
Research & scientific impact
Discovery translated into methods, models, and IP.
A scientific track record spanning genome editing, plant developmental biology, deep-learning phenotyping, and translational research.
01Patent application
Increasing periderm in plant roots
Co-inventor on a published international patent application covering compositions and genome-editing methods designed to increase root periderm and suberin.
Published international patent application · WO2023164515A2 · 2023
View application ↗02First-author transcriptomics paper
SHATTERPROOF2 cell-type transcriptomics
Connected experimental design, floral synchronization, an SHP2-specific reporter, FACS, wet-lab biology, RNA-seq, bioinformatics, interpretation, and scientific writing to resolve a reproductive meristem at cell-type resolution.
Plant Physiology · 2016 · DOI 10.1104/pp.15.01845
Read open-access paper ↗03Genome editing
From gene target to crop phenotype
Hands-on genome-editing expertise across Arabidopsis and crop systems — including cannabis and Solanaceae — connecting CRISPR strategy, molecular validation, phenotyping, genomics, and reproducible analysis.
CRISPR/Cas · Molecular validation · Crop R&D
My research has been published in
Selected publications
Contributions across discovery, genomics, and computation.
Peer-reviewed work connecting molecular mechanisms with experimental evidence, high-throughput genomics, genome editing, and quantitative methods.
Browse the complete publication record on Google Scholar ↗Plant Physiology
Transcriptomic Signature of the SHATTERPROOF2 Expression Domain Reveals the Meristematic Nature of Arabidopsis Gynoecial Medial Domain
RNA-seq · FACS · reproductive meristems · Arabidopsis
View publication ↗ScienceThe histone H3.1 variant regulates TONSOKU-mediated DNA repair during replication
Chromatin biology · DNA replication · genome stability
View publication ↗The Plant CellH3.1K27me1 maintains transcriptional silencing and genome stability by preventing GCN5-mediated histone acetylation
Epigenetics · chromatin · transcriptional silencing · genome stability
View publication ↗Nature PlantsInsight into the evolution of the Solanaceae from the parental genomes of Petunia hybrida
Comparative genomics · Solanaceae evolution · crop biology
View publication ↗DevelopmentmiR167 limits anther growth to potentiate anther dehiscence
CRISPR/Cas9 · reproductive development · Arabidopsis
View publication ↗Plant PhenomicsPAT: A Quantitative and Large-Scale Screening Method for Periderm Measurements
Deep learning · phenotyping · automated microscopy
View publication ↗Selected work
Rigorous methods. Clear decisions.
Independent examples demonstrating how I approach uncertainty, validation, reproducible analysis, and scientific communication.
01
Machine learning · synthetic data
Machine learning for biological phenotyping
Problem
Biological classifiers can look accurate while failing on new experimental groups.
Approach
Built a confidentiality-safe workflow with group-aware validation, calibration, model comparison, and interpretable features.
Result
A reproducible demonstration that exposes leakage and quantifies predictive uncertainty.
Decision
Whether a model is reliable enough to guide further biological validation.
02
Healthcare operations · data analysis
Data science for hospital operations
Problem
Delays between discharge orders and departures are important but easy to describe without testing.
Approach
Defined measurable intervals, compared groups statistically, and built decision-focused visualisations.
Result
A reproducible, patient-data-free case study separating patterns from unsupported assumptions.
Decision
Where operational review or targeted data collection should begin.
03
First-author research · Plant Physiology
SHATTERPROOF2 cell-type transcriptomics
Problem
The identity of a small reproductive domain was obscured by mixed tissue and processing effects.
Approach
Combined floral synchronization, an SHP2-specific reporter, FACS, RNA-seq, and stringent filtering.
Result
Identified 363 differentially expressed genes and evidence of a meristematic identity.
Decision
Which candidate regulators warranted functional follow-up.
04
ML phenotyping · GWAS-ready
PAT: from stained roots to quantitative traits
Problem
Manual periderm measurements limited experimental scale and reproducibility.
Approach
Combined staining, automated microscopy, and UNet++ segmentation across 20 natural accessions.
Result
Automated measurements reached 94% efficiency relative to expert measurements.
Decision
Enabled scalable traits for genome-wide association mapping and mutant screening.
How I can help
Scientific depth. Analytical discipline. Useful outcomes.
01Experimental strategy, before you run
Review the design, sampling, controls, endpoints, and analysis plan before valuable samples or resources are committed.
Explore serviceClose details+
Experimental strategy, before you run
- Review plant studies, screens, field trials, panel studies, molecular experiments, and other high-value experimental runs.
- Translate the biological question into measurable endpoints, replication, controls, quality checks, and a defensible statistical plan.
- Identify risks early and provide concise recommendations that can be incorporated before the work begins.
02Analysis when the data is already there
Turn an existing biological dataset into a documented analysis, reproducible code, clear figures, and conclusions your team can use.
Explore serviceClose details+
Analysis when the data is already there
- Assess the dataset, metadata, structure, and analytical question to define a focused route through the work.
- Clean, integrate, model, visualize, and document the data using transparent R or Python workflows.
- Deliver reusable code, publication-ready figures, and an uncertainty-aware account of what the evidence does and does not support.
03Genomics, phenotyping & predictive modelling
Build fit-for-purpose analytical and predictive workflows for genomics, imaging, phenotyping, sensory, and other biological data.
Explore serviceClose details+
Genomics, phenotyping & predictive modelling
- Work across RNA-seq, molecular assays, image analysis, longitudinal experiments, repeated measures, and multivariate datasets.
- Develop predictive workflows with feature engineering, group-aware validation, appropriate metrics, and interpretable model comparisons.
- Connect biological context to reproducible delivery through R/Python, quality control, testing, and clear documentation.
04Molecular biology & experimental troubleshooting
Independent scientific feedback for refining, troubleshooting, and interpreting molecular-biology experiments and workflows.
Explore serviceClose details+
Molecular biology & experimental troubleshooting
- Review RNA extraction, PCR and qPCR, Western blotting, cloning, CRISPR workflows, genotyping, and molecular validation strategies.
- Connect protocol choices, controls, sample quality, and downstream analysis to the biological question.
- Support experimental interpretation and next-step planning across plant molecular biology, genomics, genome editing, and phenotyping.
Experimental designData cleaning & QCMixed-effects modelsMachine learningScientific RAGFigures & reporting
Research background
Built in the lab.
Refined in the data.
My perspective comes from working across discovery science, genomics, bioengineering, and commercial R&D. I understand the experimental context behind the dataset — and the decisions waiting on the other side.
2025 — 2026
Puregene AG
Senior Scientist / Data Scientist
Cannabis genome-editing platform · applied crop R&D · quantitative analysis · genomics
2019 — 2024
Salk Institute for Biological Studies
Research Scientist
Plant genomics · RNA-seq · GWAS · machine learning · gene editing · patent application
Research group · Busch Lab
2016 — 2018
Yale University
Postdoctoral Researcher
Genome stability · epigenetics · histone modification · CRISPR/Cas9 · DNA replication
Research group · Jacob Lab
2014 — 2016
North Carolina State University
NSF Postdoctoral Fellow
Cell-type transcriptomics · FACS · developmental biology
Research group · Franks Lab
2008 — 2014
Cornell University
Ph.D., Plant Sciences
Molecular biology · genomics/transcriptomics
Research group · Mattson Lab
Project enquiries
Have a dataset or scientific bottleneck?
Send a short description of the question, dataset, or project stage. I'll reply with whether I can help and what a focused first step could look like.