The model
Overview¶
4D-Bio maps the day-to-day skills of a practicing bioinformatician onto a single framework for working with generative AI. This page explains how the model is built and how to read the four capability pages.
How the model is built¶
4D-Bio combines three existing frameworks:
| Framework | What it contributes |
|---|---|
| Anthropic's 4D framework for AI fluency | The top-level structure: four capabilities (Delegation, Description, Discernment, Diligence) that together describe fluent, responsible AI use. |
| ISCB Competency Framework v3.0 | The established, profession-wide bioinformatics competencies. 4D-Bio links each practice area to the specific ISCB competencies it draws on, so nothing is invented from scratch. |
| ACTIVE Framework | AI-specific pedagogy: guidance on task-AI alignment, verification confidence, and countering cognitive offloading. |
The 4D framework gives the shape, ISCB grounds the model in the wider profession, and ACTIVE adds the teaching-and-learning perspective for AI.
How to read a capability page¶
The four pages (Delegation, Description, Discernment, and Diligence) are organized the same way:
- A short definition of the capability, in general terms and then for bioinformatics.
- A "Practice areas on this page" box listing the capability's numbered practice areas: the professional situations where it shows up in real work.
- One table per practice area. Each row is a competency, grouped by category in the first column, with a Source tag in the last column.
Source tags¶
The tag in the Source column shows where each competency comes from:
| Tag | Meaning |
|---|---|
| ISCB | Drawn from the ISCB Competency Framework v3.0. The tag links to the competency's ISCB entry. |
| ACTIVE | Drawn from the ACTIVE framework for AI pedagogy. |
| 4D-Bio | Original to this practice model, and not yet catalogued in an external framework. |
The four capabilities at a glance¶
| Capability | Core question | Focus |
|---|---|---|
| Delegation | Should AI do this, and how? | Framing the problem, decomposing it into tasks, choosing tools, orchestrating agentic workflows. |
| Description | Did I communicate the goal well enough to reproduce it? | Translating biology into precise instructions, structuring data as context, logging provenance. |
| Discernment | Can I trust this output? | Judging biological plausibility, benchmarking against baselines, catching hallucinations and artifacts. |
| Diligence | Did I do it correctly? | Governance and compliance, data longevity, preserving manual skill, mentoring and community stewardship. |
Status¶
4D-Bio is an early public draft (v0.1). Content, structure, and terminology are all still open for discussion. See How to contribute.