Aarhus Universitets segl

Research Data Management

Practical guidance for managing your data well — from designing your experiment to archiving your findings.

Well-organized data is easier to analyze, harder to misread or lose, and simpler for others to reproduce. When data is FAIR from the start, much less effort is wasted on preparation and reformatting, and more goes to actual research. FAIR data also travels better between collaborators and earns credit when reused. Both data management plans (DMPs) and FAIR data compliance are now increasingly required by funders as conditions in grant agreements and publication policies.

This page includes the FAIR principles, a self-assessment checklist, a DMP brainstorming tool, metadata templates, and storage guides. For university-wide policies and support, see the AU Research Data Management page.

§1

The FAIR principles

Findable, Accessible, Interoperable, Reusable. FAIR is a widely adopted set of guidelines for how research data and associated metadata should be structured, described, and shared to ensure long-term usability — by all those who collect and interact with the data. Select a principle below for its formal definition and specific examples.

Findable

Data must carry a globally unique and persistent identifier (such as a DOI) and be described by rich metadata that explicitly includes that same identifier. Both the data and its metadata must be registered or indexed in a searchable resource so it can be discovered using standard search tools — for chemistry this means chemistry-specific repositories, and ideally also general scientific databases where applicable.

Examples A natural products group deposits confirmed compounds in the Chemotion Repository with InChIKeys and linked DOIs, including collection site and voucher details, so other chemists can find the structures and spectra by substructure search. A physical chemistry group assigns a DOI to every dataset and records pump wavelength, probe range, and solvent details in the README, so keyword searches on Figshare retrieve the raw data alongside the paper. A structural biology group deposits a protein structure in the PDB and NMR shifts in the BMRB with cross-references, so searching for the protein returns both the assignments and the route to raw spectra. A materials group files crystal structures in the ICSD with perovskite and composition tags, so searching by material or deposition method finds the structure and linked characterization.

Accessible

Once located, data and its metadata must be retrievable through a standard method that anyone can use, without special software or a paid subscription — in practice, a normal web link (HTTP/HTTPS) that any browser can open. The protocol can include authentication and authorization where needed — accessible does not mean open: restricted data can still be FAIR if the access route is clearly described and the metadata remains findable. Metadata must also remain accessible even when the data itself is no longer available, which is why depositing metadata in a repository early matters.

Examples An organic group deposits processed spectra openly in nmrshiftdb2 while raw instrument files stay on an internal server, with a contact email in the metadata record so others can request the raw data directly. A physical chemistry group deposits both the raw transient absorption matrices and summary traces directly on Zenodo, so everything is downloadable with no request process needed. A structural biology group deposits integrated SAXS curves openly in SASBDB while raw images stay on group storage, with the access route documented (contact email) for collaborators who need them. A materials group posts raw diffraction data and analysis scripts on GitHub, archived via Zenodo for a DOI, so the full dataset is retrievable directly with no separate request needed.

Interoperable

Data and metadata must use formal, accessible, shared languages and formats for knowledge representation so that they can be combined and understood by both people and machines. This means using community standard file formats and FAIR vocabularies or controlled ontologies where they exist — for example, InChI and SMILES for structures, CIF for crystallography, and IUPAC nomenclature. Data should also include qualified references to other (meta)data, linking your work to the broader chemical record so it can be reused across disciplines, not just within chemistry.

Examples A natural products group exports reactions as CSV with InChIKey columns and NMR data as JCAMP-DX, so other researchers can parse spectra directly without vendor software. A physical chemistry group exports transient absorption data as ASCII CSV with a JSON metadata file using SMILES (a structure code, not a stable identifier) and InChIKey (a searchable identifier, not a structure code), so others can load the data and reproduce the analysis without proprietary software. A structural biology group releases a binding dataset as CSV with SMILES (structure code), InChIKey (searchable identifier), and PDB ID columns, so computational chemists can load ligands directly into docking software. A materials group maintains a synthesis log as JSON with structured composition and linked CIF file references, so other groups can parse the file automatically and merge it with existing databases.

Reusable

Data and metadata must be richly described with accurate, relevant attributes so they can actually be reused. That means the metadata documents the full history of how the data was produced — how it was generated, processed, and calibrated, not just what version it is — so an external researcher can understand and replicate the work. It should also meet domain-relevant community standards so it can be combined with other datasets directly. A well-documented synthesis route should remain usable even separated from the paper it accompanied. Data must also be released with a clear and accessible usage license so others know what they are allowed to do with it.

Examples A synthesis group deposits NMR and HRMS files on Zenodo with batch numbers and processing details in the README, so another researcher can reproduce the spectra. A physical chemistry group publishes transient absorption data with pump wavelength, probe range, and kinetic model in the README, so a reviewer can load the raw matrix and test alternative models. A structural biology group deposits NMR titration data on Zenodo with buffer composition, pulse sequence, and software version in the README, so another group can reproduce the binding isotherm. A materials group publishes synthesis records and linked PXRD patterns on Zenodo with temperature profiles and calibration standards in the README, so a catalysis group can follow the same route and compare results directly.

§2

The Data Lifecycle

Research data passes through distinct stages from project planning to after publication. Use this section to explore the different stages and find helpful tools.

FAIR Data STAGE 1 Plan STAGE 2 Collect / Produce STAGE 3 Work / Analyze STAGE 4 Archive STAGE 5 Publish / Share STAGE 6 Reuse

Click on a stage to explore.

Stage 1

Plan

At this stage you're scoping the project's data before any of it exists — searching existing literature and repositories for data to build on or compare against, then setting up the data ecosystem the project will run on for its full duration: drafting a data management plan, assigning roles, creating project folders on shared storage, choosing an ELN, and drafting SOPs so the whole team works the same way from day one. Many groups loop in a data specialist at this point to help set conventions that keep the data FAIR-compliant from the start.

Stage 2

Collect / Produce

This is the stage where raw data actually gets generated — running experiments, writing scripts, and documenting methodology in your ELN as you go, capturing conditions, parameters, and run settings rather than reconstructing them later. You're following the SOPs and data management plan set up in Stage 1, filling in metadata templates so each dataset stays complete, and keeping files organized in the agreed folder structure on active storage so nothing gets lost mid-project.

Stage 3

Work / Analyze

Here you're turning raw outputs into results — running calculations, processing data with analysis scripts, and sharing draft figures with co-authors or collaborators. Every analysis step, parameter choice, and software version is captured as structured metadata — in the file header, a metadata file, or your ELN entry linking back to the file — so the results can be reproduced later, by the team or by your future self. As the analysis goes through iterations, each version is recorded rather than overwritten — via the naming scheme's vN suffix or a tool like git. The derived datasets and intermediate files behind your conclusions are curated and kept linked back to the original raw data, rather than left as disconnected working files.

Stage 4

Archive

This is where you step back and decide what actually needs to survive long-term — raw data, scripts, key intermediate files, and documentation — separating it from temporary files and duplicates that don't. Run the FAIR self-checklist here to guide what gets curated. Verified copies move into long-term storage, set permanently read-only, distinct from the public repository deposit that happens in Stage 5: archiving secures a complete copy first, publishing releases a curated subset later. If you're leaving AU or moving to a different group (end of PhD, postdoc, or contract), this is also when ELN ownership and access rights get handed over to your PI, before your account is deactivated — a data specialist can help with curating the collection or arranging that handover.

Stage 5

Publish / Share

At this stage the dataset moves from your control into the wider record — deposited in a trusted repository to get a persistent identifier (DOI), released under a license that states exactly how others may reuse it. The DOI gets added to your manuscript, and if release needs to wait — until after publication, or for IP reasons — an embargo delays it (for genuinely sensitive data, restricted access is used instead, since an embargo only delays release rather than controlling who gets access once it lifts). Methodology and supplementary materials go out alongside the data itself, so others can replicate or extend the work with confidence. See the repository table for where to deposit by data type.

Stage 6

Reuse

Publication isn't the end of a dataset's life — this stage is about what happens to it afterward. You cite the dataset DOI in your own reference list (e.g. Author(s) (Year). Dataset title. Repository. doi.org/xx.xxxx/xxxxxxx), and track how others build on it — most repositories (Zenodo, and other DataCite-based platforms) show download and citation counts directly on the dataset's page, so checking back periodically works better than relying on manual literature searches. Repository records — contact details, version history, documentation — get kept current, and incoming feedback or corrections get evaluated so the dataset keeps improving rather than going stale — a data specialist can help if a reuse question needs more than what's in the record.

§3

Is my data FAIR?

A fast self-check for whether FAIR practices are actually happening — not a repeat of the DMP's planning questions. Useful if you're new to FAIR and want a quick gut-check, or if you've already written a DMP and want to confirm it's being followed in practice. Once your data has a DOI, you can check how FAIR the deposit actually is with F-UJI — it scores the dataset's metadata and landing page against the FAIR principles.

0 / 11 complete

Findable

Interoperable

Accessible

Reusable

§4

Build a data management plan

This tool includes content from the European Commission's Horizon Europe DMP template, with chemistry-specific guidance added where relevant. Check your funder's requirements first — they may require a different format.

Once completed, a DMP should be stored in a shared location so all collaborators can refer to it when necessary.

Step 1 of 11

Project overview

Data summary

None selected yet

File and folder organization

Click an item below to make it the anchor for the next folder/subfolder. The tree below is generated automatically — you can also edit it by hand.

Making data findable

Making data accessible

Check your funder's open access policy before answering below — it may dictate the embargo length, license, or repository you're required to use, which affects where you can publish and how you handle the data.

Metadata accessibility

Interoperable

Reusable

Outputs and resources

Security and ethics

Your DMP is ready

Don't forget: Ask your department data manager to review this draft. Once complete, share it with all project collaborators so everyone understands the data management workflow.

§5

Practical resources

Tools, templates, storage guidance, repositories, and support contacts.

Narrow the list with the filters below: Data sensitivity — whether your data includes personal, confidential, or otherwise restricted information. Recommended data volume — an approximate guide to how much data you have, not a hard limit. Project stage — whether you're actively working with the data or moving a finished project into long-term storage.

Storage optionWhat it isBest forCost (as of August 2026)Compute
EMCCAU-hosted HPC cluster for structural biology, with GPU workstations and specialist software (RELION, cryoSPARC, AlphaFold3) accessed remotely.Cryo-EM and protein crystallography; molecular dynamics; AI/LLM; structural chemistry and biophysics — no commercial/industrial use; non-AU researchers need an active collaboration agreement with an AU faculty member.Pilot: free. Storage Only: 8,000 DKK/yr + 4,000 DKK per extra 10 TB. CPU+Storage: 12,000 DKK/yr + 6,000 DKK per extra 10 TB. Full Project: 50,000 DKK/yr; extra storage priced on request. All excl. VAT
ERDAAU's file archive for open research data, accessed with your AU login and shareable read-only or read/write.Open research data without personal or sensitive information — suitable for active use with modest data volumes during analysis, as well as datasets ready for sharing or long-term archivalFree
GenomeDKISO 27001-certified shared HPC suitable for sensitive data, with remote-desktop or job-queue access.Sensitive research data requiring GDPR compliance; compute-intensive work across all sciences; interactive analysis with GenomeDK DesktopStorage 250 DKK/TB/yr. Optional backup 500 DKK/TB/yr. CPU 0.13 DKK/hr. GPU 6–20 DKK/hr. (AU Nat/Tech/Health billed internally.)
GrendelAU's communal compute cluster for research calculations, free for modest use and active work only — move finished results to long-term storage when the project ends.General compute-intensive research; CPU and GPU calculations across sciences (the website lists no discipline-specific recommendations) — active/working storage only, not for archiving.Free up to 5 TB / 800,000 files. Chemistry group account for modest compute; larger allocations can be arranged through a PI hardware contribution.
O DriveShared folder on the AU network (appears as a drive letter), ordered via the AU service portal and accessed on the AU network or through VPN.Sharing files and documents with colleagues and students; group data storage~650 or ~360 DKK/TB/yr, depending on server. Contact AU IT to confirm your group's rate.
OneDriveCloud storage via your AU Microsoft 365 account that syncs across your devices automatically.Personal files synced across devices; easy sharing with othersFree for AU affiliates (policy may change)
SIFExtra-secure storage for personal, patient-related, or grant-restricted research data, requiring two-factor login and a short security training.Research data containing personal data or requiring extra protection — including IP, strategic, export-controlled, or audit-logged datasets — requires two-factor login and a short security training before use.Free
U CloudISO 27001-certified cloud access to a supercomputer shared by SDU, AAU, and AU, for GDPR-compliant analysis of sensitive data or computations too demanding for a personal computer.Computations too demanding for a personal computer; GDPR-compliant data analysis; sensitive data analyses; interactive work with customizable appsFree for AU affiliates and students; Tier 1 pricing on request
ELNDescriptionRecommended forPricing
AU Lab NotebookGeneral-purpose university ELN for everyday documentation.All chemistry groups — standard optionFree for AU affiliates (institutional license)
AU SciformationUniversity-hosted ELN; integrated with AU login and standard data exports.Organic / synthetic chemistryFree for AU affiliates (institutional license)
Physical or handwritten notebookTraditional bound notebook. Use only if an ELN is not feasible — digitize entries when sharing or archiving.Last resortFree (materials cost only)

Both ELNs above are AU-hosted and free for AU affiliates. Keep research data on AU-managed systems — a public or third-party ELN platform is not an appropriate place for it. Only consider an external tool if neither AU option can capture the data and metadata your workflow requires, and then only a self-hosted one with standard exports. A paper notebook is a last resort — digitize entries when sharing or archiving.

Rolling an ELN out across a group takes planning whichever tool you land on. NFDI4Chem's Choose the right ELN guide covers assessing your needs, trialling against real lab workflows, and planning training and rollout; its ELN Finder tool compares options in detail if the AU tools genuinely can't cover your workflow.

Metadata must always be as structured as possible — predefined fields filled the same way for every run, not free text. Keep it in one location close to the data (file header, a metadata file in the same folder, or an ELN entry linking to the file), and agree a file naming scheme with your group, e.g. YYYYMMDD_SampleID_Technique_Operator_vN.ext.

How to determine what metadata to record

  • Reproducibility: What would another researcher need to reproduce this exactly — every setting, condition, and protocol choice?
  • Accountability: Who was responsible, which version was used, and what changed from the SOP or last run?
  • Defensibility: If challenged in a review, patent, or audit, what evidence trail proves this result is trustworthy?
  • Lineage: Where did this come from? Link to parent datasets, raw inputs, and earlier iterations.
  • Over-record: Not sure whether to record something? It is better to have data and not need it than to need it and not have it.

Not sure which ontology or controlled vocabulary to use for a field? NFDI4Chem's Terminology Service lets you browse and search 18 curated chemistry ontologies (including ChEBI and ChMO) to find the right term.

Looking for the community metadata standard or file format for a data type? FAIRsharing is a curated registry of metadata standards, file formats, databases, and data policies you can search by discipline.

Always record the identifier in your ELN. The identifier is the permanent link between your lab record and the published dataset.

Data typeWhere to depositWhat you receiveFormats
Chromatography, electrochemistry, elemental analysis, EPR / ESR, IR spectroscopy, surface analysis, thermal analysis (DSC, TGA, DTA), UV-Vis / Raman / fluorescence, X-ray microscopy / micro-CT, general / mixedZenodoDOI on releaseAny format
Code / analysis scriptsGitHubZenodoDOI per releaseAny format
Computational chemistryioChem-BD, NOMAD, Materials Cloud, or ZenodoDOICML (ioChem-BD); VASP, QE, CP2K, etc. (NOMAD); AiiDA archives (Materials Cloud); any format (Zenodo)
Crystal structures (free deposit, no CSD/ICSD license required)Joint CCDC/FIZ Access Structures ServiceDOICIF
Crystallography (inorganic / materials)ICSDICSD code and DOICIF
Crystallography (organic / metal-organic)CSDCSD # and DOICIF, SHELX
Environmental / atmospheric dataPANGAEADOITabular, netCDF, various environmental formats
Enzyme kinetics / functional enzymology dataSTRENDA DBDOIEnzymeML (in development)
General / multidisciplinary research dataFigshareDOIAny format; 20 GB per-file limit on free accounts (institutional Figshare accounts raise this to 5 TB)
General / multidisciplinary research data (EU alternative to Zenodo)EUDAT B2SHAREPersistent identifier (handle/DOI via Invenio RDM)Any format
General chemistry research dataRADARDOIAny format
General chemistry research data (chemistry-specific instance)RADAR4ChemDOIAny format (format recommendations provided); free for researchers at German publicly funded institutions, up to 10 GB per project
Liquid-state & solid-state NMRZenodo, nmrXiv, or BMRBDOI or repo IDAny format (Zenodo); Bruker / JEOL / NMReDATA (nmrXiv); NMR-STAR (BMRB)
Mass spectrometryZenodo or MassBankDOI or MassBank IDAny format (Zenodo); MassBank text records (MassBank)
Optical & electron microscopyZenodo or BioImage ArchiveDOI or accession #Any format (Zenodo); OME-TIFF, TIFF, PNG, JPEG (BioImage)
Protein / biomolecular structuresPDBPDB ID and DOImmCIF (preferred), PDB
Scattering / diffractionZenodoDOI on releaseAny format
Supramolecular chemistry / host–guest binding dataSupraBankDOIJSON, CDX, PNG
Synthesis / reaction dataChemotion Repository (chemistry-specific); ELN; ZenodoInChI / DOIChemotion native (Chemotion); ELN-dependent; any format (Zenodo)

Don't see your data type above? Search re3data, a registry of research data repositories across all disciplines, to find a certified option for your field.

Research Data Management chatbot

A chatbot specialized in research data management, here to take the administrative burden off your plate — quick answers on formats, storage, ELN choices, metadata, repositories, and the FAIR self-checklist, plus help creating or editing your DMP.

Open chatbot →

Only reachable on the AU network — off campus, connect through AU VPN first.

Department Data Specialist

Book a consultation for DMPs, data curation, analysis workflows, or other project-level topics, or call/text/email for questions on formats, storage, metadata, or repositories.

Contact librarian

Copyright and licensing for methodology, code, data, physical materials, or technology, plus reference management, author rights, journal policies, and publishing strategy.

Contact librarian →

AU Library data management

University-wide guidance on DMPs, FAIR data, and repositories, plus training sessions and hands-on data labs (R, Python, data visualization) run by AU Library.

Visit the AU Library page →

Research Data Management for Chemists (self-paced course)

An 11-episode workshop, free and self-paced (~200 min total): general research data management, FAIR, storage, and licensing principles first, then chemistry-specific topics — ELNs, chemical data standards, and repositories.

Start the course →

Data protection coordinator (GDPR)

Personal data (participant names, contact details, health records), data protection impact assessments, or GDPR compliance questions for your project — take these to your data protection coordinator.

AU data protection contacts →

IT support

O Drive orders, OneDrive, SIF access, network issues, and general IT infrastructure.

support.au.dk →