LIMS vs ELN vs CDS vs SDMS: Laboratory System Roles and Boundaries

To select information systems for a pharmaceutical laboratory, start with the sample journey and the decisions that must remain reconstructable. A LIMS may organise samples and tests; an ELN document the activity; a CDS acquire and process chromatography data; and an SDMS collect scientific data and metadata and make them retrievable. These functions can overlap. The four acronyms do not require four purchases, nor do they demonstrate that the workflow is suitable for GMP use.

This guide proposes a functional view, an original matrix and a simulated QC case. The boundaries described are analytical criteria, not regulatory definitions of commercial categories. Each solution needs assessment of its capabilities, configuration, interfaces and intended use.

1. Map the workflow before the systems

Follow a sample from receipt to record retention: identification, test assignment, preparation, acquisition, calculations, review, approval and later retrieval. For each step, note who acts, what information enters, what record is created and what status changes. Include deviations, authorised repeat activities, incomplete results and interruptions: they belong to the process that the system must support.

A useful question is: “If I need to reconstruct this decision tomorrow, which records will I need and where are they?” The final numerical result may require the sample identifier, method and version, preparation, instrument data, processing and review. A summary screen is not necessarily the complete record set. EU GMP Chapter 4 distinguishes instructions from records and calls for electronic and hybrid records to be defined and controlled. [2]

GuideGxP recommends a simple activity map before a technology diagram. If the team cannot agree who approves a specification or which method version was effective, connecting two applications will not solve the problem. Clarify the responsibility first, then decide how the systems will represent it and how it will be verified.

2. LIMS: coordinate samples, tests and results

A Laboratory Information Management System is often the management point for the sample journey. It can assign identifiers, record receipt and status, organise tests, methods and specifications, allocate activities and collect results. It may support calculations, review and reports, according to the functions purchased, configured and verified. Do not assume that all these capabilities are available simply because the product is called a LIMS.

When describing its role, separate master data from transactions. An approved specification differs from a result assessed against that specification; a method version differs from the test performed on a sample. Clarify which objects the LIMS manages, who maintains them and how the test retains its reference to the version used. Otherwise, an update may make the meaning of a historical result ambiguous.

The LIMS can receive a result from the CDS without becoming the system that retains every original chromatography record. It can hold a link to an ELN activity without containing every attachment. These boundaries are acceptable only when declared and supported by access and review workflows. “Result present” does not automatically mean “complete evidence available”.

3. ELN: document execution and context

An Electronic Laboratory Notebook can record protocols, preparations, observations, calculations and links to activity data. In QC, its value is more than replacing a sheet of paper with a digital page: it maintains the context needed to understand what was done, by whom and with which materials and instructions. Controlled templates can help, provided they do not prevent the recording of an unexpected event.

Distinguish a protocol from its execution. An approved template may be reused; the record for a sample must preserve the version applied and what actually occurred. Corrections, attachments and reasons need management within the activity’s context. A prepopulated field does not demonstrate that a step was performed, and a signature cannot recover observations that were never recorded.

The EPFL institutional experience reported by Argento highlights the importance of organisational support and change management in ELN/LIMS implementation. It is an academic experience, not evidence of GMP validation. [5] For pharmaceutical laboratories, GuideGxP therefore proposes testing real activities and roles before assuming that interface flexibility or familiarity equates to fitness for use.

4. CDS: acquire and process chromatography data

A Chromatography Data System typically manages chromatography acquisition and processing and, where supported, instrument control. Sequences, acquisition and processing methods, signals, integrations, results and change information contribute to reconstructing the analysis. The boundary with firmware, the instrument and other systems depends on the architecture; it needs description rather than inference from the software’s name.

A result transferred to the LIMS may be a number with a unit and review status, while the dynamic record remains in the CDS or an appropriate archive. A summary PDF does not necessarily preserve the original’s information and ability to be interrogated. FDA guidance distinguishes static from dynamic records and explains the role of metadata in reconstructing the activity. [3]

In the project, identify which processing produced the transmitted value and how the reviewer reaches the relevant records. If authorised reprocessing changes the result, the workflow must make the version, reason and consequences for the previously received data recognisable. Avoid both silently overwriting the earlier value and retaining two apparently definitive results without a relationship.

5. SDMS: collect, organise and retrieve

A Scientific Data Management System can collect files and metadata from instruments and applications, organise and index them, and make them searchable. Acquisition, version management and retention capabilities depend on the solution. The term SDMS does not guarantee support for every format, dependency or function needed to read the data over time.

Define what is transferred: reports alone, complete folders, native data, metadata, relevant audit trails or linked record sets. Check how the system recognises a complete transfer and signals a missing item. A list of filenames does not show that the content can be retrieved with its original meaning. Relationships between records also need to survive retention.

Also separate long-term retention from operational recovery. MHRA distinguishes an archive from a recovery backup; a temporary backup does not automatically replace retention of final records. [4] Agree terminology with IT and the applicable requirements: FDA also uses “backup” in the context of the retained copy required by CGMP. [3] Specify the purpose and retrieval test, not just the label.

6. Assign authority and interface responsibilities

The authoritative source is the one designated by the process for a particular item of information and stage. It does not necessarily mean one system for the entire laboratory or the only copy in existence. The LIMS may govern sample status and the CDS the chromatography record; a retained copy must maintain the relationship and characteristics needed for its use.

For every exchange, GuideGxP proposes documenting the object, identifier, version, unit, permitted status and person responsible for reconciliation. Clarify what “sent”, “received”, “accepted” and “available for review” mean. An active connection does not prove that a record was accepted into the correct process. A value without a unit, or assigned to the wrong sample, can cross a network without a technical error.

EU GMP calls for appropriate checks on electronic exchanges and an up-to-date description of data flows and interfaces for critical systems. Annex 11 also addresses roles, risk management, validation and the lifecycle. [1] The following matrix turns these questions into a working example; it is not an architecture prescribed by the Annex.

7. Original activity–record–control matrix

In this possible configuration, the laboratory uses four functions, which could also be delivered by fewer platforms. “Responsible system” identifies where the function is managed; people’s responsibilities still need assignment. The choices must be confirmed against the laboratory’s requirements.

Use the matrix to find cells without an owner and duplicated functions. If two systems calculate the same result, document which calculation governs the decision and how differences are handled. If no system retains a necessary element, that is a gap to resolve: an arrow in a diagram does not close it.

8. Simulated case: from receipt to approved result

A QC laboratory receives a sample for a chromatography test. In this simulated case, the LIMS assigns its identifier and test with the applicable method and specification. The analyst opens a linked ELN activity, records preparation and observations, and uses the same reference in the CDS sequence. No real data or universal acceptance values are introduced.

The CDS acquires the data and retains the information needed to reconstruct processing. After the specified review, it sends the result, unit, identifier and reference to the processed version to the LIMS. The workflow checks the match and status before making the data available for the next decision. The ELN remains accessible to explain preparation, without recopying its context into a free-text LIMS field.

During a project test, the laboratory simulates a result received without its link to the preparation activity. The proposed workflow holds it as incomplete and assigns the exception to a person. Correction restores the link with evidence of the intervention; it does not create a second, apparently independent record. This tests the ability to reconstruct the activity, not just transfer speed.

The reviewer consults the record set required by the procedure, assesses exceptions and records the decision. Approval of a QC result is not presented as batch certification or release. The SDMS collects the sets defined by the strategy, and the team tests retrieval while preserving identifiers and relationships. If interrogating the data requires the source system, that dependency remains explicit and managed.

This is one possibility. Another laboratory might document preparation using a controlled LIMS function or adopt a different retention arrangement. The choice is judged against workflow completeness, control and retrievability, not the number of applications installed.

9. Set priorities and check dependencies

Compare the current and required workflows. Identify the problems first: untraceable samples, transcription, incomplete context, fragmented review or difficult retrieval. For each, define a necessary function and a test demonstrating its effectiveness. A commercial demonstration using perfect data does not cover the laboratory’s exceptions.

Ask to see a method change, an incomplete transmission, an unauthorised access attempt and retrieval of a historical record with its attachments. These are GuideGxP proposals to adapt to risk. Also assess skills, interface maintenance, continuity, export and service dependencies. A useful function may require configurations or roles that the laboratory has not yet established.

Assign QC, system owners, IT and quality responsibilities for the process, data, support and changes. Avoid leaving the integrator as the only person who understands a field’s meaning. Plan intended-use validation and change control with proportionate evidence. No “GMP-ready” label, electronic signature or cloud platform automatically makes the entire activity compliant.

Operational conclusion

The architecture is understandable when the laboratory can identify where each record originates, which function manages it, how it passes to other systems, and how it is reviewed and retained. Start with the sample, assign responsibilities and test exceptions. Acronyms provide orientation, but the decision needs evidence about the configuration actually used.

Continue in the Digital Lab & Data Integrity HUB and explore the essential principles of Data Integrity in QC laboratories.

Sources and applicability

Sources verified on 30 September 2026. Annex 11 and Chapter 4 are the current published EudraLex texts for the human-medicine framework discussed here; consultation proposals are not treated as effective. FDA and MHRA provide guidance in their respective contexts. The EPFL experience is academic and was consulted through its abstract and indexed excerpts, not in full. The matrix, case and selection scenarios are original GuideGxP proposals, not obligations to purchase particular systems.

1. European Commission — EU GMP Annex 11: Computerised Systems. Revision 1, effective 30 June 2011.
2. European Commission — EU GMP Chapter 4: Documentation. January 2011 revision.
3. FDA — Data Integrity and Compliance With Drug CGMP: Questions and Answers. Final guidance, December 2018.
4. MHRA — GXP Data Integrity Guidance and Definitions. Revision 1, March 2018; for UK GLP, consider the precedence of OECD 2021 stated by MHRA.
5. Argento — Institutional ELN/LIMS deployment. EMBO Reports, 2020;21:e49862. DOI: 10.15252/embr.201949862.