IT systems · Practical guide

Validating Cytiva UNICORN: linking the method, run and analysis

In Cytiva UNICORN, an apparently plausible result can omit the settings changed during a run or use a different evaluation. Trace instrument control, in-run changes and the analysis actually used. QA, laboratory, production and IT teams can then determine which supplier evidence transfers and what they need to test locally.

Table of contents

At a glance

What does Cytiva UNICORN validation involve?

UNICORN validation demonstrates that the software, configuration and workflows in use support their intended purpose. Link instrument, method, actual run, authorised interventions and analysis to requirements and tests. Assess the relevance of supplier evidence. Software functionality, instrument suitability and suitability of the analytical method require different evidence. Actual use determines the scope and depth.

  • Device, software version, instrument configuration, and licenses determine the scope.
  • Method development, in-process testing, and subsequent analysis will be evaluated separately.
  • A correct calculation does not establish an appropriate measurement chain or analytical method.
  • Supplier records, planned outages, and operational controls provide the evidence.

UNICORN Version, Device, and Licenses Jointly Determined

Cytiva UNICORN integrates instrument control, method processing, data analysis, and administration. The name alone does not fully describe your installation. An ÄKTA workstation with a local database has different dependencies compared to multiple instruments using centralized data storage. Capture device, software version, instrument configuration, database, user management, and additional licenses in a single, unified overview.

This article draws on the UNICORN 7 data sheets dated 15 September 2025 and 24 June 2026. The product page also lists licence items for version 7.13 and later. This does not identify the latest patch. For your installation, obtain the relevant compatibility matrix and change information.

Evaluation Classic, Column Logbook and interface functions may require additional licences. UNICORN start for ÄKTA start is a separate product. Check that an observed workflow or supplier evidence applies to the specific package. Record which functions your laboratory uses to define a meaningful scope for testing.

Sources: [1] [3] [2]

Software, device, and method have different validation objectives

Start with the determination of the UNICORN scope: Is a cleaning procedure being developed, a manufacturing step being controlled, or a result for a quality decision being evaluated? A research run does not automatically carry the same regulatory weight as a GMP process, even if it bears the same product name. Also specify whether and for what purpose its data will be used subsequently.

The regulatory framework for this article is Human Medicinal Product-GMP. Annex 11 is a GMP guideline for computerised systems. Other industries and a potential FDA/Part 11 reference will be determined separately based on actual record-keeping obligations and usage.

The software validation addresses, for example, whether a parameter is correctly captured and documented. Device suitability encompasses, among other aspects, the appropriate measurement range and verification of the measurement chain. For an analytical application, the suitability of the testing method is additionally considered. A mathematically correct peak analysis does not validate either the method's selectivity or the accurate performance of the hardware. Link these evidences together, without repeating the same objective under three separate headings.

Sources: [4] [5] [9] [10] [8]

Establish a verifiable traceability pathway from the modules

Use the four modules to assign requirements. Start by describing the expected result. “Method Editor is available” records what is installed. “The planned run uses the approved wash phase for the correct column” is testable.

The following work aid assigns internal audit questions to documented modules. It is not a complete manufacturer's functionality list. For each row, add the responsible person, existing evidence, and pending review. This makes it visible whether a requirement is already adequately covered or only refers to a general product specification.

Internal Work Aid: Linking Requirements and UNICORN Module
ModuleQuestion to askAppropriate comparison
Method Editor Which method with which parameters may commence? Approved instruction for prepped run
System Control Which settings actually take effect during the run? Default, permitted intervention and time course
Evaluation Which evaluation bears the professional responsibility? Run dates, evaluation status, and results
Administration Who is authorized to make changes and how are data retained? Task-specific Rights and Recovery Documentation

Sources: [2] [5]

Consider method specification and actual operation separately.

The general specification sheet describes methods from phases as well as watch instructions that respond to a signal condition. Therefore, for process evaluation, start values alone are not sufficient. The condition, its action, and a permissible manual intervention may all influence the subsequent course. Whether a specific function is compatible with the device should be verified against the device's documentation.

Cytiva describes parameter changes during routine operation and their capture in the Run Log. This enables the derivation of a dedicated inspection focus: Do the recorded operator actions, the subsequently effective settings, and the actually observed process behavior align? A static, unaltered method design does not answer this question.

Determine which changes are technically permissible and who is authorized to perform them. Prepare a valid and an invalid trial. The second trial shall demonstrate whether the intended control is effective. Watch and alarm functions are process-related features; their presence does not constitute functional safety or universal protective capability for any system.

Sources: [2] [3] [5]

Distinguish process run evaluation from method suitability assessment

The UNICORN Evaluation supports the assessment of results, peaks, and fractions; the Evaluation Classic expands the range of analysis capabilities. Therefore, the specific calculation or processing required must be determined prior to test planning. An additional license is neither a quality assurance measure nor automatically included in every installation.

Your workflow should clearly define which data feed into the decision-making process: original run curves, relevant interventions, selected analysis approach, and, if applicable, fraction assignments. A reviewer must be able to understand why this particular analysis method was selected. Establish clear procedures for changes to ensure that prior and subsequent evaluations remain traceable within the existing system.

Verify calculations using independently defined data and set values. This control demonstrates the implementation of a computational rule. It does not replace experimental evaluation of an analytical method. Similarly, a plausible curve profile and correct software output do not serve as evidence that a column is suitable for the product or that a sensor has been properly calibrated.

Sources: [2] [10] [9] [6]

Further reading:

Sample calculation: Wash volume and altered flow rate

The example is neither a manufacturing instruction nor a completed UNICORN test. For the example method M-02, the column volume is 20 ml; the wash-in instruction specifies five column volumes, abbreviated as CV. This results in 5 × 20 ml = 100 ml. The original flow rate is 2 ml/min.

For pure theoretical calculation, we assume a constant actual flow rate, immediate transitions, and no additional delays. Without intervention, 100 / 2 = 50 minutes would be required. An allowable change to 4 ml/min after 20 minutes yields 40 ml initially, followed by 60 ml in 15 minutes: a total of 100 ml in 35 minutes. This ideal calculation does not reflect real pump accuracy or a specific UNICORN stop condition.

Before testing, record method M-02, column reference C-20, the assigned roles and the starting conditions. Define how the volume setting will be implemented, how volume will be measured and which tolerances apply. Start each case with a new test identifier and reset parameters. Retain the evidence from each run. Use a suitable test setup without a production batch.

Own Expected Table: calculated expectations, no observed test results
Failure and EntryIndependently established expected resultQuestion the evidence should answer
Normal: M-02, C-20, 5 CV, 2 ml/min 100 ml; ideal 50 min Voices associated method, column and course?
Permitted flow change after 20 min to 4 ml/min 40 ml + 60 ml; ideal 35 minutes total Are the intervention, timing, and further course distinguishable?
Accidentally entered old instruction with 4 CV Only 80 ml instead of 100 ml Will an incorrect input be identified prior to professional acceptance?
False column cover with 25 ml and 5 CV 125 ml instead of 100 ml Is an inappropriate assignment identified and resolved?
User without change permission tries to switch flows No effective change made by this user Does the permission defined for this test work?
Report shows unchanged 2 ml/min despite authorized intervention The actual change must be identifiable during evaluation. Is a misleading validation evidence identified?

Sources: [5] [6] [2]

Review records according to their meaning in the process

For the example run, simply finding a login entry somewhere is insufficient. The technical review requires a clear linkage between the instruction, the run, the intervention, and the applied evaluation. Determine which information is required in the original system and in the report. A custom-designed summary of results may be useful, without necessarily including all required source data.

Assign permissions based on tasks: method preparation, conducting productive runs, evaluation, technical review, and administration. In a small organization, roles may be combined; however, the associated risks and effective controls must remain clearly understandable. Specifically assess whether a routine account has the capability to modify relevant settings that fall outside its assigned responsibilities.

A user acknowledgment is not automatically considered an electronic signature. If an electronic signature process is prescribed or required, the meaning, assignment, and authorizations are to be separately defined. Even in the presence of existing technical functionalities, the question remains open as to whether the specific workflow complies with current requirements.

Sources: [5] [6] [1]

Obtain supplier documentation in accordance with applicable requirements.

Cytiva offers, among other things, Validation Support Files as well as IQ/OQ support. These offerings represent a sensible starting point. However, the actual package available is decisive: Which version, device connectivity, configuration, and requirements does each individual validation document address? A document title alone does not answer that question.

Assign a document its verified function, test environment, results, and open deviations. If it meets your requirements and transmission conditions are satisfied, the same test does not need to be repeated solely due to a different author. Conversely, a comprehensive supplier audit does not close gaps that lie outside the scope described by the supplier.

In the example, appropriate documentation could cover technical rights management. Whether your roles are correctly assigned and the employed cleaning method is compatible with the column basis remains to be specifically determined. Plan additional assessments for such open questions. The number of documents or screen shots is not an adequate substitute for this alignment.

Sources: [2] [5]

Maintain data retention and changes during operational activities

The industrial data sheet differentiates between database responsibilities of a SQL Express installation provided as part of the package and a central SQL installation. Therefore, it is essential to clarify who is actually responsible for backup and restoration. The statement "IT handles this" is only actionable when database volume, ownership responsibilities, and a traceable restoration procedure are clearly defined.

Select a run with relevant change and associated evaluation for your own recovery case. After recovery, the information should be traceable, readable and technically related. In case of external data retention, check whether local and passed records represent the same run. An existing OPC connection does not confirm a correct allocation to the receiver.

Evaluate updates, device modifications, new analyses, and extensions based on their impact. The vendor overview also highlights beta extensions and potential support limitations. A new extension incurs its review requirements based on its usage. Consolidate relevant requirements, existing documentation, additional assessments, and the decision regarding continued use.

Sources: [3] [5]

Further reading:

Preparation and participation within a small team

The process owner defines the intended operation and technical boundaries. The laboratory or production responsible party provides methods and credible data. IT evaluates architecture and data retention; QA assesses requirements and validation evidence. The manufacturer or integrator supports in addressing version and device-specific limitations. It is not necessary for a single individual to provide all of these contributions.

The effort increases significantly, particularly with differing methods, different devices, individually used extensions, and unclear data transfers. Existing relevant documentation and accessible test environments facilitate the work. The size of the organization alone does not justify a reduced level of inspection; what matters most are risk, complexity, and existing knowledge.

  • Device, UNICORN-Build, Instrument Configuration, database and licenses are specified.
  • Intended use, methods, relevant evaluations, and external recipients are described.
  • Permissible interventions and task-related rights have been clarified.
  • Existing documents are assigned according to request and version fit.
  • Test data, independent setpoints, initial state, and reset are prepared.
  • Technical review, restoration, and decision on open items are the responsibility of the responsible party.

Sources: [5] [7]

Project experience and collaboration

Assess the method version used for the UNICORN run

To interpret a recorded UNICORN run, you need its method and the version used to evaluate the results. With your process owners, we trace that link and examine the supporting evidence. For a method change, we identify the steps affected. For a new application, we first establish the purpose, modules used and decisions made. Your team can then choose tests appropriate to the intended use.

Frequently asked questions about Cytiva UNICORN

Has UNICORN been validated by Cytiva for our GMP use?

Manufacturer documentation may provide appropriate evidence for the scope described. For your intended use, additional alignment is required for device, version, methods, data pathways, and operational tasks. Take appropriate documentation with justification and address remaining questions. A manufacturer's description of regulatory support or an existing IQ/OQ package does not constitute a general validation of your entire process.

Sources: [2] [5]

Which UNICORN version is referenced in this guideline?

The guideline utilizes the UNICORN-7 product sheets from September 2025 and June 2026, along with the product page featuring 7.13-related license positions. It does not claim to represent the most recent patch level. For your project, you require the actually installed build, the associated compatibility matrix, device configuration, and relevant changes. Similar-named packages do not constitute evidence of identical functional scope.

Sources: [1] [3] [2]

Do we need Evaluation Classic for a GMP application?

GMP compliance alone does not establish a requirement for Evaluation Classic. The extension is optional; what is decisive is the specific analyses required and the functionalities of the actually licensed modules. Verify that the selected analysis is technically appropriate, used correctly, and enables a traceable and defensible decision. The package name does not substitute this evaluation.

Sources: [2] [10]

May operators change parameters during a run?

Only to the extent that the approved process anticipates such interventions and the user has the necessary authorization. The manufacturer's documentation outlines technical intervention capabilities. Define the permitted scope and establish criteria for assessing the impact of such interventions. Conduct a test run with an authorized and a non-authorized attempt. A run protocol supports traceability but does not, by itself, permit arbitrary process changes.

Sources: [2] [5]

Does a correct UNICORN calculation demonstrate the suitability of our method?

We initially validate only the verified calculation step under the considered conditions. For an analytical method, additional subject-matter expertise and supporting evidence are required. Equipment measurement, sample handling, and data evaluation each have distinct potential sources of error. An appropriate comparative calculation can identify computational errors, but pure software review cannot confirm selectivity, actual measurement accuracy, or process suitability.

Sources: [10] [9]

Is a results report sufficient for review?

The report must clearly demonstrate the actions taken, data references, and evaluation statuses relevant to your decision. Where such information is missing, additional records from the original system may be required. Explicitly establish and verify this linkage during an actual review case. The PDF format does not guarantee completeness nor immutable retention.

Sources: [6] [5]

How do we handle new UNICORN-extensions?

Capture purpose, origin, version, support status, and impact on your data or process control. The manufacturer notes that beta extensions may also be available. This does not imply automatic suitability for your use. Changes will be evaluated based on the affected requirements; existing documentation and targeted additional assessments will be consolidated accordingly.

Sources: [3] [5]

What must be available prior to the start of the validation activity?

The most helpful elements are a clear installation overview, the methods and analyses employed, and a described data flow. Supplement this with defined roles, existing supplier documentation, and appropriate access for the planned studies. Additionally, identify who determines intended reference values and evaluates deviations. While documents can be initiated without these inputs, no credible acceptance decisions can be made.

Sources: [5] [3]

What determines effort and duration in a small and medium-sized enterprise?

Critical factors include variations and open questions regarding equipment, methods, data analysis, interfaces, and the quality of existing documentation. The availability of qualified personnel and testing capabilities also influences planning. The management of a single, clearly defined work station differs from that of multiple individually linked systems. Relying solely on a fixed timeframe—such as "UNICORN" or employee headcount—is therefore not sustainable.

Does an update automatically trigger a complete revalidation?

First, changes and their impacts are assessed. Affected areas may include device connectivity, permissions, reporting, or recovery procedures. This assessment determines the required validation evidence. A justified selection of additional checks may be sufficient; a significant change may require more comprehensive work. The decision should clearly indicate what remains unchanged and which new risks have actually been evaluated.

Sources: [5] [3]

Sources and review date

As of 6 October 2026. Product specifications are based on UNICORN-7 data sheets from 15.09.2025 and 24.06.2026 as well as the product page. Subversion, device and license scope are to be compared in the project; 7.13 is not described as the latest.

  1. Cytiva UNICORN 7 control software
  2. UNICORN 7 software — System control software, CY12681
  3. UNICORN 7 software — Control software for industrial systems, CY50315
  4. European Commission: EudraLex Volume 4
  5. EU GMP Annex 11: Computerised Systems
  6. EU GMP Chapter 4: Documentation
  7. EMA: ICH Q9(R1), Step 5, Corr.2
  8. FDA: Part 11 — Scope and Application
  9. EU-GMP Chapter 3: Facilities and Equipment
  10. EU-GMP Chapter 6: Quality Control

Sources checked: .

Daniel Herrmann, Daniel Herrmann Consulting

Daniel Herrmann Consulting

Carry out the validation of your UNICORN deployment

We support you in distinguishing device, method, and data transfers and in compiling appropriate evidence for your intended use.

  • Computer system validation →

    Bring your GxP systems into validated operation, from laboratory software, MES and eQMS to ERP. We define the scope, plan risk-based tests using GAMP 5 and hand the evidence to QA for release.

Send a project enquiry Customer testimonials →