What survives?
Measure residual tumor burden and identify surviving clones or lineages.
Residual Disease Intelligence
Understanding what survives cancer therapy.
AustinBioLabs develops computational approaches to study molecular residual disease, drug-tolerant persistence, and the evolutionary paths that lead to recurrence.
Multi-omicsFunctional genomicsEvolutionTarget discovery
Therapy
Residual disease
Persistence
Resistance
AustinBioLabsmaps the transition
The problem
Cancer treatment can remove most tumor cells while leaving behind a small population capable of surviving, adapting, and eventually driving relapse.
Residual cells are not necessarily smaller versions of the original tumor. They can occupy distinct transcriptional, regulatory, metabolic, and signaling states that create new dependencies—and potentially new treatment opportunities.
Research thesis
Move beyond “MRD positive” toward a biological model of residual disease.
What state does it enter?
Resolve residual-cell transcriptional, chromatin, signaling, and microenvironmental states.
What does it depend on?
Use functional genomics to distinguish state markers from causal survival dependencies.
How can it be eliminated?
Prioritize interventions designed to target the residual state before overt resistance develops.
Platform direction
Residual Disease Intelligence
A computational framework that links molecular state, functional dependency, evolutionary trajectory, and therapeutic opportunity.
Inputs
RNA / scRNA
ATAC / chromatin
CRISPR screens
Tumor genomics
Longitudinal ctDNA
Residual State Enginestate · dependency · evolution
Outputs
Residual State Score
Dependency ranking
Regulatory circuits
Resistance hypotheses
Target prioritization
Flagship research program
EGFR Persist
Mapping the osimertinib drug-tolerant persister state in EGFR-mutant non-small-cell lung cancer using public multi-omic and functional-genomics datasets.
From broad response to a compact residual-state model.
The first analyses integrate RNA-seq, ATAC-seq, H3K27ac, functional annotations, and treatment-state trajectories to separate acute response from established persistence.
S0Parental
→
S1Acute
→
S2Persister
→
S3Washout
→
S4Recovery
20compact Core genes in the current public-data model
3therapeutic evidence classes: dependency, state regulation, resensitization
4EGFR-mutant NSCLC model backgrounds in the discovery analysis
1continuous Residual State Score designed for external validation
State
Quiescence is active biology.
The persister state is characterized by suppression of proliferative E2F/MYC/FOXM1 programs together with broader regulatory remodeling.
Chromatin
Persistence is regulatory reprogramming.
RNA and accessibility changes indicate that drug tolerance involves a reconstructed gene-regulatory state, not only inhibition of EGFR signaling.
Function
Expression is not dependency.
The next layer integrates CRISPRn/CRISPRa screens to distinguish resistance drivers, sensitizers, and plausible therapy-pressure survival requirements.
Research displayed here is based on public datasets and computational analyses. It is exploratory research, not a clinical diagnostic or treatment recommendation.
Residual-state atlas
One signature, multiple biological stories.
The compact DTP-Core-20 signature is shown as induced and repressed programs, then connected to selected target hypotheses and pathway shifts across S0-S4.
20DTP-Core genes
16,606integrated gene universe
188pathway programs
5S-stage states
Core-20 induced
Core-20 repressed
Added target nodes
Selected evidence
Surface biomarker
ICAM1
Baseline
High
Repressed
Mixed
Pathway selection by stage
Residual state
S2 Persister
Activated programs
Suppressed programs
Methods
Evidence should converge before a target becomes a hypothesis worth testing.
01State specificity
Separate established persistence from acute drug response and generic stress.
02Cross-model conservation
Require reproducibility across independent EGFR-mutant contexts.
03Regulatory support
Connect RNA changes with chromatin accessibility, enhancer activity, and TF programs.
04Functional evidence
Integrate CRISPR and perturbation screens rather than equating expression with dependency.
05External validation
Test frozen signatures in independent studies, patient-derived systems, and single-cell data.
Long-term vision
A general map of therapy-induced vulnerabilities.
EGFR/osimertinib is the first model system. The broader goal is to learn which residual states recur across targeted therapies, which dependencies are created by treatment, and which interventions can prevent those states from becoming resistant disease.
Founder
Austin George
Computational Biologist
AustinBioLabs is an independent research and platform concept focused on computational cancer biology, molecular residual disease, and functional genomics.
austingeorge123@gmail.com ↗Collaborate
Working on residual disease, functional genomics, or targeted therapy?
Academic and translational collaborations are welcome, especially projects involving multi-omic treatment response, CRISPR screens, longitudinal molecular data, and target prioritization.
Contact Austin