Research · Five Domains. One Substrate.

Where we apply
the compute.

terminalBio's unified molecular intelligence engine — BindFM, XUZU, RNJ — is being directed at the hardest open problems across five research domains. One compute substrate. Many frontiers.

Explore Models Our Mission →
01
Research Area

Cancer
Research.

Targeting oncoproteins, predicting drug resistance, and designing novel cancer-binding aptamers — all within one unified molecular compute framework.

Cancer is, at its most fundamental level, a molecular recognition failure — misfolded proteins escape immune surveillance, oncoproteins hijack signaling cascades, and cancer cells evolve resistance by altering binding interfaces faster than conventional drugs can follow.

terminalBio's approach applies BindFM's cross-modal binding prediction to the specific challenge of oncology: predicting how mutant oncoproteins differ in their binding behavior from wild-type, mapping the full interaction surface of cancer driver mutations, and identifying de novo small-molecule or aptamer inhibitors using XUZU's generative engine.

Resistance emerges at the binding interface. If you can model every binding interaction as one continuous function, you can predict resistance mutations before they appear in the clinic.

The RNJ model — which achieved Kaggle Silver (89th of 1867 globally) for RNA 3D structure prediction — directly applies here: oncogenic RNA folds, miRNA-mRNA interactions in cancer gene regulation, and the structural basis of aptamer-based cancer targeting are all tractable within the same framework that solved competitive RNA folding.

Current focus areas include: KRAS G12C binding pocket modeling, aptamer design against HER2 and PD-L1 extracellular domains, and RNA-based regulatory network mapping in triple-negative breast cancer.

BindFM XUZU RNJ Aptamer Therapeutics PPI Inhibition
Key Applications
Oncoprotein Binding Prediction
BindFM predicts affinity shifts caused by driver mutations in KRAS, p53, BCR-ABL, and EGFR against known inhibitor scaffolds.
Aptamer-Based Cancer Targeting
XUZU designs RNA and DNA aptamers against tumor-associated surface antigens without experimental SELEX — 100 candidates per target, ranked by predicted Kd.
RNA Oncology Circuits
RNJ's 3D RNA structure prediction enables modeling of oncogenic miRNA hairpin structures and lncRNA regulatory interactions in cancer transcriptomes.
Resistance Mutation Scanning
Systematic mutagenesis of binding interfaces to predict which variants confer resistance before clinical emergence — enabling pre-emptive combination therapy design.
PD-L1 / Immune Checkpoint Design
Designing novel nanobody and aptamer binders for PD-L1 and CTLA-4 immune checkpoints — with cross-modal affinity heads trained on antibody-antigen structural data.
02
Research Area

Drug Design &
Development.

The entire small-molecule drug discovery pipeline — from target identification to lead optimization — reimagined on a single cross-modal compute substrate.

Drug discovery is a binding problem. Every step in the pipeline — hit identification, lead optimization, selectivity profiling, ADMET prediction — reduces to predicting or engineering molecular recognition events across the protein-ligand interface.

BindFM was designed with this pipeline in mind. Its protein-small molecule binding head is trained on binding affinity datasets across multiple assay types (Kd, Ki, IC50, kon/koff), giving it sensitivity to the kinetic dimensions of drug binding that static docking programs miss.

A drug doesn't just bind — it binds at a rate, for a duration, with a selectivity profile. The next generation of binding models must predict all three simultaneously.

Beyond prediction, XUZU's generative OT-CFM (optimal transport conditional flow matching) decoder can be directed toward small-molecule generation conditioned on a target binding pocket — enabling fragment-growing and scaffold-hopping without enumerating chemical space explicitly.

Cross-modality is critical here: aptamers increasingly compete with small molecules in therapeutic applications. XUZU enables the same design loop — target structure in, candidate binder out — across both chemical and nucleic acid space, from one interface.

BindFM XUZU Affinity Prediction Lead Optimization Virtual Screening
Key Applications
Virtual Screening at Scale
BindFM's shared encoder processes protein-ligand pairs without re-encoding the target for each compound — enabling ultra-fast virtual screening across large chemical libraries.
Kinetic Profiling (kon / koff)
Predicting binding kinetics alongside equilibrium affinity — residence time is a key determinant of in vivo efficacy that most binding models ignore entirely.
Nucleic Acid Therapeutics
XUZU extends the drug design paradigm to RNA and DNA aptamers, siRNA molecules, and antisense oligonucleotides — modalities that can reach previously undruggable targets.
Selectivity Profiling
Predicting off-target binding across homologous protein families (kinomes, GPCRs, nuclear receptors) from a single BindFM model — no separate per-family fine-tuning required.
Covalent Inhibitor Design
Modeling the covalent warhead reaction coordinate as an affinity prediction problem, enabling rational design of irreversible inhibitors against cysteine- and serine-containing binding sites.
03
Research Area

Proteomics &
Genomics.

From protein interaction networks to CRISPR guide design, TB Biocompute treats the proteome and genome as one molecular system — not two separate databases.

The proteome and the genome are not independent entities. They are the same molecular system observed at different levels of abstraction. Proteins bind DNA at transcription factor binding sites, RNA polymerase reads DNA templates, ribosomes translate mRNA — every cellular process is a series of nucleic acid-protein binding events.

BindFM's protein-nucleic acid binding head was designed precisely for this cross-domain landscape. The same weights that predict aptamer-protein affinity also predict DNA-binding domain interactions, transcription factor occupancy, and CRISPR-Cas9 PAM recognition events.

The proteome and genome are the same physics problem at different scales. A unified binding model doesn't need to be told which level it's operating at — it reads the atoms.

In genomics applications, RNJ's RNA 3D structure prediction capability maps directly to riboswitch function prediction, guide RNA secondary structure validation, and splicing regulatory element identification — all of which depend critically on RNA fold accuracy.

For proteomics, BindFM's protein-protein interaction head enables interactome-scale binding prediction — predicting which protein pairs interact, at what affinity, and how disease mutations alter the interaction network. This is the foundation for systems-level drug target discovery.

BindFM RNJ CRISPR PPI Networks Splicing
Key Applications
CRISPR Guide RNA Design
Predicting guide RNA on-target cleavage efficiency and off-target binding across the genome using BindFM's nucleic-nucleic protein binding head — no separate fine-tuning per Cas variant.
Transcription Factor Binding
Predicting TF-DNA binding affinity as a function of both DNA sequence and TF structural conformation — capturing cooperativity and chromatin context effects.
Protein-Protein Interactome
Systematic prediction of the human protein-protein interaction network with per-pair affinity estimates — enabling identification of high-confidence drug targets in disease interactomes.
RNA Splicing Regulation
RNJ models the structural basis of exonic splicing enhancers and silencers — predicting how mutations at splicing regulatory elements alter pre-mRNA secondary structure and RBP binding.
Ribosome Translation Efficiency
mRNA secondary structure directly regulates ribosome engagement and translation rate. RNJ's 3D fold predictions enable quantitative modeling of codon-structure effects on protein expression.
04
Research Area

Computational
Biology.

The architectural research that makes everything else possible — novel geometric deep learning, equivariant representation theory, and the design of the TB Biocompute substrate itself.

terminalBio is, at its core, a computational biology research organization. The models — BindFM, XUZU, RNJ — are as much research contributions as they are tools. Each architectural decision is a hypothesis about the geometry of molecular space.

The central hypothesis: SE(3)-equivariant graph neural networks operating directly on atomic coordinates are the correct inductive bias for any molecular prediction task. Not protein language models, not evolutionary sequence statistics — the 3D atomic graph.

The architecture is the hypothesis. EGNN shared weights across protein and RNA carbons are not an engineering shortcut — they are a claim about what binding is, at the physical level.

Active architectural research includes: scaling laws for equivariant molecular models (how does prediction accuracy scale with parameter count and training set size?), the design of FEBI (Fundamental Encoding Block Intelligence) and RJ (Relational Junction) — XUZU's proprietary transformer and graph attention blocks — and the GER (Generate-Evaluate-Refine) self-improvement loop that uses REINFORCE to push generated sequences toward lower predicted Kd.

TB Biocompute is the abstraction layer above the model weights — the API, evaluation harness, and model composition framework that allows models trained on different binding types to be composed into multi-step prediction pipelines.

EGNN SE(3)-Equivariance FEBI Blocks OT-CFM GER Loop
Research Directions
Equivariant Molecular Scaling Laws
Empirically characterizing how prediction accuracy, generalization, and OOD robustness scale with EGNN depth, embedding dimension, and training data size across modalities.
Cross-Modal Transfer Learning
Does pretraining on protein-protein binding data improve RNA-protein binding prediction? Systematically characterizing transfer learning dynamics across the five binding modalities.
FEBI & RJ Architecture Research
Ablation and scaling studies of the Fundamental Encoding Block Intelligence transformer and the Relational Junction graph attention mechanism — the proprietary blocks at XUZU's core.
Generative Model Theory (OT-CFM)
Optimal transport conditional flow matching for discrete molecular sequence generation — theoretical and empirical analysis of convergence, mode coverage, and Kd-conditioned generation quality.
TB Biocompute OS Design
The API and composition framework that makes terminalBio's models composable — designing the evaluation harness, model versioning protocol, and multi-head prediction pipeline abstraction.
05
Research Area

Aging.

Aging is a molecular accumulation problem. Protein aggregation, epigenetic drift, senescence signaling — all reducible to binding interactions that become pathological over time.

Aging biology is, at the molecular level, a story about binding events going wrong slowly. Proteins misfold and aggregate. DNA repair machinery loses fidelity. Epigenetic marks drift from their regulatory positions. Senescent cells sustain SASP (Senescence-Associated Secretory Phenotype) signaling by altering their cytokine binding landscape.

Every one of these processes is a change in molecular binding behavior — and therefore addressable with a sufficiently general binding prediction framework.

Aging is entropy accumulating at binding interfaces. If you can model every binding interaction, you can model aging — and intervene at the molecular level before the phenotype becomes irreversible.

BindFM applies to aging research in several ways: predicting how amyloid beta and tau protein aggregation changes PPI networks in neurodegeneration, modeling how sirtuins and NAD+-binding enzymes change their kinetics as NAD+ levels decline with age, and predicting which senolytics disrupt senescence-sustaining PPI interactions.

XUZU's aptamer design capability is particularly relevant here: aptamers against aggregation-prone proteins (alpha-synuclein, amyloid beta, tau) represent a growing therapeutic modality for neurodegenerative aging pathologies. The zero-shot design capability — no SELEX required — dramatically lowers the experimental barrier to testing new candidates.

RNJ contributes through telomerase RNA structure modeling and the prediction of age-associated changes in RNA secondary structure stability — a component of the RNA degradation that accompanies cellular aging.

BindFM XUZU RNJ Senolytics Neurodegeneration Epigenetics
Key Applications
Amyloid Aggregation Modeling
Predicting how amyloid beta and alpha-synuclein change their protein-protein interaction profile during aggregation — identifying intervention points accessible to small molecules or aptamers.
Senolytic Target Design
Predicting which BCL-2 family protein-protein interactions sustain senescent cell survival, and designing specific inhibitors that selectively eliminate senescent cells while sparing healthy tissue.
NAD+ Pathway Modeling
Modeling how sirtuin-NAD+ binding kinetics change with declining NAD+ availability — identifying which sirtuin isoforms are most sensitive to supplementation strategies.
Anti-Aggregation Aptamer Design
XUZU designs RNA and DNA aptamers targeting aggregation-prone protein epitopes — with zero-shot candidates ranked by predicted Kd against the monomeric target.
Telomerase RNA Structure
RNJ models the 3D structure of telomerase RNA (TR/TERC) and its protein subunit interaction surface — enabling rational design of telomerase modulators for both aging and oncology applications.