Centre Guidelines
Mission, expectations, roles and reporting, minimum commitment, communication, meetings, source code, data management and reproducibility.
Read the policy →A volunteer centre asks people to work without pay. The least it owes them in return is to write down what it expects, what they get, and what happens when either side falls short. Those documents are public, and so is every change made to them.
CHIRAL is the research arm of the DeepBio ecosystem, alongside DeepBio Limited, which provides bioinformatics and data services as a company, and DeepBio Academy, which runs training. The three are governed separately because membership means different things in each: employment, volunteering, and enrolment respectively.
That separation matters for reading anything on this page. CHIRAL's rules — unpaid positions, minimum commitment, authorship conditions — are research-specific. They do not govern DeepBio Limited employees or Academy trainees, and are not applied to them.
Every position at CHIRAL, Research Assistant and Group Leader alike, is voluntary and unpaid. It is a research and training opportunity, not employment. Nobody is offered a salary, a stipend, or any financial benefit.
In return the centre provides mentorship, training, access to infrastructure, co-authorship on work a member genuinely contributed to, and — on satisfactory completion — a recommendation letter. Because the position is unpaid, those four things are also the only levers the centre has if commitments are not met. They are earned, and they can be withheld.
The Director leads the centre. Each unit is led by a Group Leader who reports to the Director, and Research Assistants work within a unit under their Group Leader. Every member knows who they report to and what their role is expected to produce.
Centre-wide policy takes precedence over any unit's own overlay. A unit may add rules for its own tools, data types and stack; where the two conflict, the centre document wins unless the Director approves a documented exception.
Entry runs through a prerequisite course, then an interview — not through who you know. A candidate completes the DeepBio Academy course required for the unit they are applying to, applies with proof of completion and a written disclosure of outside commitments, interviews with the unit's Group Leader, and is selected by the Group Leader with the Director's approval. The first three months are probationary.
Completing the course makes someone eligible, not selected — selection is competitive. The required course, the proof, the interview outcome and the decision are recorded for every recruit, so selection can be examined after the fact.
Authorship follows the four ICMJE criteria, all of which must be met. Gift, guest and ghost authorship are prohibited. Roles are discussed at the start of a project and revisited as contributions change, each author's specific contribution is recorded using CRediT, and seniority does not buy a place on a paper.
This is the mechanism behind the claim made elsewhere on this site that students first-author their own work. It is written down so it can be held against us.
Missed revisions, unexcused absences and missed communication windows follow a graded ladder rather than arbitrary judgement. A first lapse is a logged reminder; a second is a written warning with a catch-up plan; a third triggers an authorship review against the ICMJE criteria; persistent or severe lapses mean removal from the project and referral to the Director. Penalties affect only the member who lapsed — they never reduce the credit of people who did the work.
Research misconduct, disputes over authorship, and appeals each have their own written procedure in the Publication & Authorship Policy.
No identifiable human data is collected or analysed without ethics approval, and data is used only for the purpose the participant consented to. Sensitive data lives in encrypted, access-controlled storage — never personal laptops, personal cloud drives or personal inboxes — with access granted by name on a least-privilege basis and revoked at offboarding. Identifiers are separated from analysis datasets, precise household coordinates are treated as identifying, and no identifiable data or credentials go into Git.
They are living documents, kept in the same public repository as this website. Anyone can propose a change by opening a pull request, and every revision carries an author, a date and a rationale in the history. What the centre required of a member in a given month can be reconstructed exactly.
These are the documents a member signs up to at onboarding. Each links to the canonical copy in the repository, with its full revision history.
Mission, expectations, roles and reporting, minimum commitment, communication, meetings, source code, data management and reproducibility.
Read the policy →ICMJE authorship criteria, author order, CRediT contributor roles, revision responsibilities, the accountability ladder, dispute resolution and misconduct.
Read the policy →Data classification, ethics approval and consent, storage and least-privilege access, de-identification, sharing agreements, and breach handling for health data.
Read the policy →How Research Assistants and Group Leaders are selected: the prerequisite course for each unit, interview, selection, probation, and the records kept of each decision.
Read the policy →Coding standards, reproducibility, documentation and version control — what an analysis has to satisfy before it leaves the centre.
Read the policy →How members represent the centre online, including the affiliation requirement and the escalation that applies when it is not met.
Read the policy →