BIOMEDICAL DATA SCIENCE CONSULTING
Oncology bioinformatics, statistical modeling and translational research. We work as an external data science department for your R&D team: you bring the research question, we deliver figures and conclusions ready for the decision meeting and the data room.
WHO WE WORK WITH
THE PROBLEMS WE SOLVE
You need to decide which target to advance first, or which indication to pursue next. We use public data from TCGA, ICGC, GEO and CCLE to show where a target is overexpressed in tumor versus normal tissue, where the normal-tissue safety window lies, and how expression relates to immune infiltration and prognosis. You don't need to share any internal data for this.
Your scRNA-seq, bulk RNA-seq, WES or ctDNA data is waiting for analysis, or your bulk results are ambiguous because averaging across a whole tissue dilutes the signal. We integrate multi-omics data and analyze it at single-cell resolution. QC report first, then cell annotation, cluster maps and sample-level results.
Project reviews, licensing partners and journal reviewers all ask the same question: where does this number come from? Our statistical and predictive models are cross-validated and externally validated. Every number traces back to its source, and the code ships with the report.
CORE CAPABILITIES
01
Pan-cancer expression specificity, pathways, signatures and immune infiltration from TCGA, ICGC, GEO and CCLE. Hi-C / HiChIP 3D genomics to link non-coding GWAS variants to the genes they actually regulate.
Deliverablestarget/indication evidence pack, publication-grade figures, reproducible scripts
02
scRNA-seq, bulk RNA-seq, WES and ctDNA. Genome, transcriptome and proteome data integrated around the tumor microenvironment.
Deliverablescell annotation and cluster atlas, sample-level results tables, QC report
03
Survival analysis, logistic regression, GEE, GLM/Lasso; Random Forest, SVM and other machine learning methods, with cross-validation and external validation.
Deliverablesstatistical analysis plan, model code and validation, nomogram or risk stratification tool
04
Biomarkers, clinical sample data and antibody drug strategy. Design and interpretation of in vivo and in vitro studies.
Deliverablestarget assessment report, study protocol and sample size, manuscript support
PRODUCT · EVIDENCE PACK
A standard-format report on a single target that answers one question: in which indication does this target have a case, and where are the risks? Built for licensing discussions, project initiation and indication selection.
WORKFLOW
turn the R&D question into a testable hypothesis.
AI-assisted scans of literature and databases; an expert decides whether the data can answer the question.
reproducible pipelines, every number traceable to its source.
map results back to mechanism and clinical context.
figures and conclusions ready for the meeting room and the data room.
HOW WE WORK
We use AI every day: LLMs and agents help scan literature and clinical trial databases, generate code and draft reports. A computational biologist checks every number, and every conclusion traces back to its source.
SELECTED COLLABORATION
IF 29.5 (JCR 2025; subject to the journal's official figure)
PLT012, a humanized anti-CD36 antibody, to release anti-tumor immunity in liver cancer and liver metastasis. The work combined public TCGA pan-cancer, CIBERSORT, scRNA-seq and GSEA data with preclinical and human HCC ex vivo validation to support proof of concept for CD36-targeted immunotherapy. Co-authored with a Swiss clinical-stage client.
a nomogram for combination-therapy benefit in EGFR-mutant NSCLC
dynamic ctDNA neoantigen monitoring, with molecular response detected before imaging response
DATA SECURITY
Your data stays on your servers. We work through remote access.
Confidential from day one. NDA signed before kickoff; everyone on the project is bound by it.
Returned or destroyed. All copies are returned or destroyed at project close, as agreed.
Compliance first. We follow human genetic resources and data security regulations. Client data never goes into public models.
Start with a free 30-minute consultation. Walk us through your question. We'll tell you whether we can answer it, how, and what data it takes. If we can't, we'll say so.
