Collaboration: Fraunhofer SCAI × APTA Therapeutics
Scope: Transforming fragmented Post-COVID / Long COVID / PASC clinical evidence into a structured, reproducible, graph-queryable knowledge layer
Status: Active · Task-gated delivery
🌐 Live dashboard → https://postcovid-apta.netlify.app
Post-COVID research is scattered across hundreds of clinical trials, observational studies, and case reports — written in natural language, inconsistently reported, and hard to query systematically.
This framework builds a pipeline that ingests that literature and outputs a structured knowledge graph: entities (interventions, populations, endpoints, mechanisms), relationships between them, and evidence scores tied to source documents. The result is a queryable evidence layer that can answer questions like:
- Which interventions have been tested for Post-COVID conditions, and with what outcomes?
- Which patient populations and cohort characteristics are most represented?
- Which clinical endpoints are measured most consistently across trials?
- Which mechanistic hypotheses have the strongest evidential support?
- What is strong enough to inform future trial design or therapeutic prioritisation?
The project is organised into five modular tasks. Each task has its own deliverables and a go/no-go review point before the next stage begins.
| Task | Name | Deliverable |
|---|---|---|
| Task 01 | Foundation — Curated Post-COVID Study Corpus | Curated corpus · structured study catalogue · interactive dashboard |
| Task 02 | KG Schema & Common Data Model | Documented schema · entity mapping rules · harmonisation workflows |
| Task 03 | SCAIView Semantic Search (optional) | Annotated corpus · semantic search interface |
| Task 04 | Text-to-Graph Extraction | Normalised triple dataset — endpoints, cohort descriptors, temporal context |
| Task 05 | Evidence-Weighted Knowledge Graph | Neo4j KG with evidence scoring · provenance · query workflows |
APTA_PostCOVID_TextToGraph/
│
├── README.md ← this file
├── assets/
│ └── kg_preview.png ← knowledge graph preview
│
├── Task01_Curated_Corpus/
│ ├── README_Task01.md
│ ├── 01_Dashboard/ ← interactive HTML viewer
│ ├── 02_Catalogues_CSV/
│ ├── 03_Methods_Provenance/
This repository is shared under the terms of the Fraunhofer SCAI · APTA collaboration agreement. Contents are confidential and intended for project stakeholders only.
Fraunhofer Institute for Algorithms and Scientific Computing (SCAI) · Sankt Augustin, Germany