☰

DIANA Biotechnologies a.s.

Connecting single-cell screening to AI-ready data

Industry
Biotechnology
Headquarters
Vestec, Czech Republic
Public information as of
January 2026

A4BEE prepared this analysis from publicly available sources. It reflects our own reading of DIANA Biotechnologies a.s.'s published strategy and is not endorsed by, or produced in cooperation with, DIANA Biotechnologies a.s.. Company website

Strategic priorities

DIANA Biotechnologies runs four interdependent business units from a single Vestec site near Prague: Diagnostics, Life Science reagents, contract research services, and an internal monoclonal antibody programme. The 384-well DIANA assay (a DNA-linked Inhibitor Antibody Assay, a small-molecule probe linked to a DNA tag read out by qPCR, or quantitative polymerase chain reaction) is the company's scientific centrepiece, and the 2023 legal merger of DIANA Lab and DIANA Engineering into a single joint-stock company made the in-house robotics team a strategic asset rather than a side project.

The past two years have shifted the strategic centre of gravity. In 2021 DIANA reported CZK 926.8 million of revenue (roughly EUR 38.7 million) with consolidated net profit of CZK 580.3 million, and management has earmarked an EUR 20 million R&D war chest to fund the pivot from COVID diagnostics to drug discovery. The monoclonal antibody programme screens around 10,000 single-cell clones per day and is positioned to feed AI models with quantitative assay data, but only if the data is captured in a structured form rather than left in instrument files and local scripts.

Manufacturing scale and regulatory scope are also changing. Life Science reagents (PCR mixes, enzymes, RNase inhibitors, oligonucleotide synthesis) ship from the same Vestec site at industrial volume, and the European IVDR (In Vitro Diagnostic Regulation) framework now requires dynamic traceability from Master Cell Banks through fermentation, purification and formulation into the fielded kit. The compliance boundary runs through every operational system on site, from the production lines and the lab robotics to the LIMS (Laboratory Information Management System) and ELN (Electronic Lab Notebook) stacks.

The same Vestec campus houses the Department of Robotics, the chemical synthesis labs, the manufacturing lines and the screening facility. That physical co-location is a strategic advantage, and the resulting digital environment — industrial SCADA (Supervisory Control and Data Acquisition) systems next to research-grade ELNs, custom gantries speaking to commercial qPCR banks — is the place where the next phase of work will either compound or fragment.

Challenges we see

  • Operations Manufacturing

    Keeping custom screening robots running through long campaigns

    DIANA's internal robotics team builds gantry systems that move plates between dispensers and banks of qPCR instruments to support 100,000-compound screens. The team owns the full maintenance burden and operates without a 24/7 vendor support line.

    Where robots are bespoke and downtime hits mid-screen, the practical answer is monitoring the machine's own signals — motor torque, temperature, cycle counts — so a developing fault surfaces before the run does.

  • Digital Integration

    Routing 10,000 daily clones into data an AI model can read

    The single-cell screening platform produces around 10,000 clones per day, generating FACS (Fluorescence-Activated Cell Sorting) data, NGS (Next-Generation Sequencing) reads and qPCR curves that must reach the antibody optimisation models as structured inputs.

    When assay data lives in instrument files and local Python scripts, the path to a model training set runs through manual exports. Pulling raw data into a shared, lineage-tracked store changes what the data science team can build against.

  • Compliance Regulatory

    Closing the IVDR traceability loop on reagent genealogy

    Life Science reagents (PCR mixes, enzymes, RNase inhibitors) ship at industrial volume under ISO 13485. IVDR requires dynamic traceability from Master Cell Banks through fermentation, purification and formulation into the customer kit and field performance data.

    Connecting each cell bank lot, fermentation run and formulated batch to the kit it eventually becomes requires records that travel with the material, so the audit question is answered from the system rather than reconstructed for it.

  • Operations Operations

    Allocating instruments between paying clients and internal programmes

    DIANA operates as both a CRO and an internal drug developer. The same HTS (High-Throughput Screening) platforms serve kinase profiling, inhibitor screening and CRO campaigns alongside the internal oncology monoclonal antibody programme.

    Where the same instruments serve fee-for-service work and proprietary programmes, scheduling and access control move from informal rules into the system, so client data is segregated by design rather than by convention.

  • Digital Integration

    Speaking one digital language across biology and engineering

    The 2023 merger unified DIANA Lab and DIANA Engineering into a single entity, but the engineering team building robots and the biologists using them still work in different software environments: SCADA on one side, research ELNs on the other.

    Bridging the two sides with a common protocol layer (SiLA 2 (Standardisation in Lab Automation) or OPC UA (Open Platform Communications Unified Architecture)) makes a custom robot just another node in the lab's data network, rather than an island needing per-vendor drivers.

Opportunities, by urgency and business impact

Each bubble is one opportunity, numbered to match the list below. Further right means it bites sooner; higher means a bigger effect on the business. A bigger bubble means a bigger implementation effort.

Source: A4BEE analysis of public sources

  1. Sensing fatigue in custom robotics before a screen fails

    Custom-built screening robots have no vendor hotline and no standardised reliability telemetry, so the maintenance burden falls on the internal engineering team during long campaigns.

    Instrumenting motors, drives and temperature-critical axes with sensors that feed a predictive model surfaces developing faults in time to schedule maintenance around, rather than during, a screening run.

    • DIANA Biotechnologies company page, Department of Robotics recruitment signals
    • The Recursive, Biotech Innovators of CEE profile, 2025

  2. Making every clone screen produce a model-ready data point

    Single-cell screening produces around 10,000 clones per day across FACS, NGS and qPCR outputs, and management has stated that quantitative DIANA-Ab data is what improves the antibody optimisation AI model.

    A shared data lake that ingests raw instrument output, normalises it against assay metadata and exposes it to the data science team turns the screening throughput into a training set rather than a storage problem.

    • MEDICA 2025 DIANA Biotechnologies presentation, antibody development
    • DIANA, Deep Research Strategic Dossier, 16 January 2026

  3. Carrying reagent genealogy from cell bank to customer kit as data

    Life Science reagents (PCR mixes, enzymes, RNase inhibitors) ship at industrial volume under ISO 13485, and IVDR requires that each kit be traceable back through fermentation, purification and formulation.

    Capturing each step as a digital record with the material lot it acts on creates the genealogy IVDR asks for and feeds the Golden Batch analysis that links process parameters to enzymatic activity outcomes.

    • DIANA Biotechnologies, Life Science product range
    • Strategic Dossier, regulatory traceability analysis, 16 January 2026

  4. Scheduling instruments and segregating client data by design

    DIANA runs CRO campaigns alongside internal drug discovery on the same HTS platforms, and client molecule structures must remain separated from internal R&D databases throughout the work.

    A scheduling layer tied to role-based data segregation makes instrument time and data access follow project rules automatically, so the CRO/R&D boundary is enforced by the system rather than by convention.

    • Strategic Dossier, Service-Product Conflict analysis, 16 January 2026

  5. Making custom robots speak SiLA 2 or OPC UA to the rest of the lab

    Custom gantries and banks of commercial qPCR instruments (Roche, Bio-Rad) currently rely on per-vendor drivers and ad-hoc scripts to exchange data, which slows the addition of any new instrument or screen.

    Implementing SiLA 2 or OPC UA across the screening facility turns each instrument into a documented network node, so adding a new machine or screen becomes a configuration task rather than an integration project.

    • Strategic Dossier, Department of Robotics tech stack inference
    • DIANA case study, 384-well DIANA assay and HTS platform

What we'd propose

  • Digital CDMO

    Predictive maintenance overlay for DIANA's custom screening robots

    We instrument DIANA's custom robotics with sensors on motors, drives and thermal axes, stream their values into a time-series model, and surface developing faults on a dashboard the engineering team can act on before a screening run is interrupted.

    • Sensor overlay on custom robotics

      Reading the machine's own signals

      Fit motor torque, vibration, temperature and cycle-count sensors to the gantry axes and qPCR banks, streaming values through OPC UA so each axis publishes its own operating data rather than relying on what the controller happens to log.

    • Predictive model on operating history

      Forecasts before failures

      Train a model on historical maintenance events so degradation signatures (rising torque, drift in thermal profiles, climbing cycle counts on wear parts) trigger a maintenance window before a fault interrupts a screen.

    • Maintenance dashboard for the engineering team

      One view of robot health

      Build a Grafana-style dashboard showing robot health, predicted maintenance windows and alert thresholds, so the engineering team schedules interventions around campaigns rather than responding to breakdowns mid-run.

    • Maintenance is scheduled around screening campaigns, not during them.
    • Engineering effort shifts from emergency repair to planned intervention.
    • Reliability becomes a measured property of each machine rather than a feeling.

    We've done this before

    See how we do Digital CDMO
  • Enterprise AI

    Single-cell screening data lake feeding the antibody optimisation model

    We design and deliver a shared data platform that ingests FACS, NGS and qPCR output from the screening facility, applies standardised ETL (Extract, Transform, Load) under FAIR principles, and exposes lineage-tracked datasets to the antibody discovery AI programme.

    • Multi-modal instrument ingestion

      Raw data captured at source

      Build connectors that pull FACS .fcs files, NGS .fastq reads and qPCR Ct values directly from the instruments into a central store, with the assay run, plate and well identity attached as lineage metadata.

    • Cloud ETL under FAIR principles

      Structured data for the AI team

      Run cloud-native transformations that normalise assay outputs, link them to plate maps and compound libraries, and publish the result as a structured dataset ready for model training rather than as cleaned-up spreadsheets.

    • Lineage-tracked dataset access

      Traceable from instrument to model

      Expose the dataset through query and retrieval interfaces that carry full lineage back to the source run and instrument, so every training example can be traced and every model output explained.

    • The 10,000-clone-per-day screening run becomes a steady training input rather than a data archival problem.
    • Antibody optimisation models learn from data with provenance, not from reconstructed tables.
    • Future instruments and assays attach to the same ingestion pattern rather than each needing a custom pipeline.

    We've done this before

    See how we do Enterprise AI
  • Digital Lab

    Digital batch records linking reagent genealogy to fielded kits

    We replace paper batch records for the Life Science reagent line with digital capture at every step from Master Cell Bank through fermentation, purification and formulation, so IVDR traceability is produced by the line rather than compiled for it.

    • Electronic batch records for reagent manufacture

      Manufacturing as data

      Capture formulation, fermentation and purification steps as digital records with material lots, operator identity and timestamps attached, replacing the paper and static PDF batch records inherited from the COVID-era scale-up.

    • Reagent genealogy across unit operations

      Cell bank to customer kit

      Link each Master Cell Bank vial through Working Cell Bank, fermentation, purification and formulation to the kit it becomes, so the genealogy IVDR asks for is a property of the system rather than a manual exercise.

    • Golden Batch analysis on process parameters

      Best runs become the reference

      Analyse historical batch data to identify the process parameters (temperature, pH, time) that correlate with the highest enzymatic activity, and use the result as the validated reference for new batches.

    • IVDR traceability is answered from the system rather than reconstructed for the audit.
    • Reagent manufacturing moves from compiled records to produced records.
    • The same dataset that satisfies the regulator also feeds process improvement.

    We've done this before

    See how we do Digital Lab
  • Digital Lab

    Scheduling and data segregation across CRO and internal R&D

    We deploy a scheduling and access layer over the HTS platform and the underlying data stores, so CRO client work and internal drug discovery programmes share instruments without sharing data.

    • Instrument booking system

      Time on the robots, planned

      Provide a reservation layer over the screening robots and shared instruments, with rules for CRO projects, internal programmes and maintenance windows, so allocation conflicts are resolved before the run starts.

    • Role-based data segregation

      Client data stays client data

      Apply role-based access controls at the database level so a CRO client's molecule structures and assay data remain inside the client project boundary, with full audit logging of every access.

    • Utilisation analytics for management

      Capacity visible at a glance

      Surface instrument utilisation, project consumption and forecast capacity on a management dashboard, so the trade-off between CRO revenue and internal programme value is decided on numbers rather than recall.

    • Client confidentiality is enforced by the data layer rather than by procedure.
    • CRO revenue and internal R&D value are visible on the same dashboard.
    • Capacity bottlenecks surface in the booking data before they show up in missed deadlines.

    We've done this before

    See how we do Digital Lab
  • Digital CDMO

    Lab 4.0 protocol layer across custom robots and commercial instruments

    We implement SiLA 2 or OPC UA across the screening facility, so custom gantries, qPCR banks and downstream LIMS/ELN systems communicate through a documented protocol layer rather than per-vendor drivers.

    • Protocol standardisation

      One language across the lab

      Implement SiLA 2 for laboratory equipment and OPC UA for industrial automation, so every instrument in the screening facility publishes its capabilities and status through a documented interface.

    • Universal data drivers

      One driver layer, many vendors

      Build the driver layer that polls qPCR banks (Roche, Bio-Rad), the custom gantries and the LIMS, aggregating their data into a single facility view rather than requiring an engineer to check each vendor's screen.

    • Closed-loop deconvolution for pooled screens

      Hits become confirmation runs

      Wire positive hit lists from pooled screens into a deconvolution workflow that queries the compound library, identifies active constituents and generates the robot worklist for the confirmation run without manual handover.

    • Adding a new instrument becomes a configuration step rather than an integration project.
    • Custom robots and commercial instruments sit on the same data network.
    • Pooled screening hits move from spreadsheet to worklist without a manual step.

    We've done this before

    See how we do Digital CDMO

Digital maturity: today and target

Scored out of 100 across six dimensions. The target is what DIANA Biotechnologies a.s.'s own published ambition implies — not a perfect score.

Source: A4BEE analysis of public sources
Data integration 35 → 85
Screening output lives in instrument files and local scripts; the antibody optimisation AI ambition runs against data with no shared lineage or central store. The EUR 20 million R&D budget is in part an investment in the data foundation the AI programme will sit on.
Process automation 55 → 90
Custom gantries run the screening facility, but they run without condition monitoring, and pooled-screen deconvolution appears to rely on manual worklists. Closing the loop between hit identification and confirmation run is the next automation step.
Quality management 40 → 85
IVDR requires dynamic traceability from Master Cell Bank to customer kit and field performance. The QMS inherited from the COVID scale-up still leans on paper batch records and static PDFs, so the audit path is compiled rather than produced.
IT/OT convergence 30 → 80
Engineering and biology sit under one legal entity since the 2023 merger but still operate in different software environments — SCADA on the engineering side, research ELNs on the biology side. A common protocol layer is the bridge.
Analytics and AI 25 → 75
Management has stated that quantitative DIANA-Ab data is what improves the antibody optimisation model. The model exists; the data infrastructure that would feed it at the stated throughput does not yet.
Cybersecurity and governance 45 → 80
CRO client confidentiality, internal R&D IP and IVDR field performance data all converge in the same Vestec site. Role-based access and audit logging are present in parts, but the unified governance layer that covers all three is still being assembled.

Check this yourself

Our Service Portal has free self-assessments and market comparisons. These are the ones that line up with what we've read above — no sales call required.

Think we've read this right?

Talk to us

Related reading

This is an independent analysis prepared by A4BEE from publicly available information as of January 2026. It reflects A4BEE's own interpretation and opinion, is not affiliated with, endorsed by, or verified with DIANA Biotechnologies a.s., and may be incomplete or inaccurate. All company names and trademarks are the property of their respective owners. To request a correction or removal, contact [email protected].