Learning Center

Working documentation for the antibody R&D platform. Start with the platform workflow, then use module tutorials, result-interpretation guides and project references as needed.

Featured
Platform Walkthrough — Models & Methods
Technical tour of the platform as it actually runs: how a wet-lab decision is mapped to a computable question, why these models are learned solvers rather than language models, the full model inventory by decision, how models compose into workflows, and an explicit record of what is reachable in production versus merely implemented. EN / 中文.

canonical technical reference · status-honest

Open
Research x AI Workflow
External-facing Research x AI narrative for antibody R&D: public/patent recall, NGS/phage pools, semi-de-novo triage, Interface AI, affinity ranking, evidence packets, CAR-T validation, and wet-lab feedback loops.
Open
Tutorials
Current per-feature walkthroughs for the real sidebar: NGS Pool, Sequence Patent Scout, Semi-de-novo, Interface, Runs & Candidates, Affinity, G-protein Design-around, CAR Interface Beta, RFantibody, Fold, Viewer, Compare, Models, and WIP specialized entries.
Open
Case Studies
Evidence-led examples showing how platform capabilities support concrete antibody and CAR-T research decisions, from target fit through candidate review.
Open
From scFv to In-vivo CAR-T
Interactive system map: antibody recognition → full CAR → targeted LVV, seven engineering layers, and the three gates of entry, expression, and tumor action. 中文主页面。
Open
Semi-de-novo Walkthrough
Real semi-de-novo triage funnel: validate on known positives, then run a 1.2M-clone production pool toward a ~100-clone shortlist with structure-based scoring and honest limits.
Open
Interface Guide
How to read Interface jobs today: complex prediction, contacts, epitope/paratope residues, BSA, pDockQ, affinity lookup, CSV/PDB exports, and when to use Interface Jobs or Compare.
Open
Affinity Metric — Plan & Candidate Models
Affinity Engine plan and candidate model stack: Boltz-2 head, PRODIGY, AlphaBind, GearBind, DSMBind, mCSM-AB, wet-lab-driven selection rules, and fine-tuning once data lands. Bilingual EN + 中文 interleaved.
Open
RFantibody Walkthrough
How the RFantibody module maps structure-based antibody design onto the platform: target PDB, hotspots, framework choice, RFdiffusion-Ab, ProteinMPNN, RF2-Ab gates, artifacts, history, and GPU runner handoff.
Open
RFantibody CDR Boundary Contract
Residue-level scientific guide to Chothia-derived HLT generation masks versus IMGT sequence annotation, using the real h-NbBCII10 framework with exact boundary differences, insertion codes, API fields, and mobile rendering rules. EN / 中文.
Open
Readout Guide
Operating manual for the scores visible across /interface, /runs-candidates, and related review surfaces: biological meaning, calculation, reliability, ranking usage, and worked examples.
Open
Scoring Audit
Honest assessment of exposed scoring metrics: what's calibrated, what's heuristic v0, and what should change before a score becomes a hard gate.
Open
Research x AI References
Focused reference list for the Research x AI story: public data resources, display/NGS, structure/interface models, affinity/RL, CAR-T validation, and G-protein design-around.
Open
Wet-lab Loop
How computational evidence should flow back into wet-lab decisions and how assay results should update the next model or triage round.
Open
Knowledge Map
Mental model for the platform's current surfaces: format universe, sequence substrate, evidence flow, use-case matrix, and module mapping.
Open
Use Cases
"I have X. I want Y. What is the path?" Common recipes composed from current platform tools.
Open
References
Foundational papers, product-context patents, construct files, and internal sketches that ground the platform's design choices.
Open
Platform Whitepaper
Legacy/background master document for the earlier Praxis six-step narrative. Useful historical context, but not the current operating map.
Open
De Novo Workflow
Legacy/background canonical de novo binder workflow. Keep for context; current tutorials and Research x AI Workflow are the source of truth for daily work.
Open