Screening Hub logo
Academic Research · Data Visualization

Screening Hub

Screening Hub is a deep-learning platform that analyses microscopy images from cell screening campaigns in drug discovery. Its four modules cover 2D imaging, transcriptomics, 3D tissue data and high-producer cell selection, delivered as a full analysis service.

Active GDPR compliant Contact Sales No public API Verified by Guidaio
Overview

What is Screening Hub?

Screening Hub is a deep-learning platform built for cell screening, the step in drug discovery where cells are exposed to chemical agents and thousands of microscopy images are recorded and evaluated to find compounds that produce a measurable — and usually desired — cellular effect. The publisher frames the problem in figures: on average twelve years and around four billion euros per drug, a process it calls extremely expensive, lengthy and risky. Manual evaluation of microscopy images is the bottleneck it attacks, together with false hits and a chemical search space kept small by the cost of looking.

The platform is modular, and four modules are described. The Classic module runs high-content screening using an advanced Cell Painting approach: six dyes stain eight cellular components and organelles — nucleus, endoplasmic reticulum, nucleoli and RNA, F-actin cytoskeleton, Golgi, plasma membrane and mitochondria — for a detailed reading of cell morphology. The Transcriptome module combines microscopy-based high-throughput screening with transcriptomic data through multimodal clustering, which the vendor says identifies hits without genetic manipulation of the cells. The 3D module processes tissue sections directly, pairing 3D convolutional neural networks with graph neural networks. The High Producer Cells module targets biopharmaceutical manufacturing by spotting high-yield single cells; the solution page describes it as still under development.

Use follows four steps: raw screening data is uploaded, the algorithms analyse it and surface candidates, visualisations are generated, and actionable recommendations are returned. Around that sits a full service — needs analysis, data processing, evaluation and interpretation, then reporting — so that, in the vendor's words, there is no implementation effort on the customer's side. Results are scored with the Z-factor, the industry standard in high-throughput screening, and the vendor claims roughly ten percent better classification accuracy than conventional machine learning.

Named references back the application examples: Institut Pasteur Paris and a leading German pharmaceutical company in immunology, aimed analytics in oncology, Fraunhofer IGB and 420 Pharma in dermatology. The tool is published by ValueData GmbH, a German company, and is sold on individual quotation rather than as self-service software.

What it does

  • Analyse thousands of microscopy cell images automatically
  • Identify compound hits that produce a measurable cellular effect
  • Classify cell-type distribution within an image, across different cell lines
  • Integrate transcriptomic and imaging data through multimodal clustering
  • Process 3D data such as tissue sections without flattening them
  • Identify high-producer single cells for biopharmaceutical production
  • Deliver validated analysis reports and interpretable visualisations
Audience

When to use Screening Hub / When not to

A quick filter to help you decide if Screening Hub is the right fit.

When to use Screening Hub

  • Pharmaceutical and biotech discovery teams running high-content screening campaigns on compound libraries
  • Contract research organisations that need client image data analysed without building an in-house AI team
  • Academic and institute laboratories working on cell-based phenotypic assays in immunology, oncology or dermatology
  • Groups holding both transcriptomic and imaging data who want the two read together rather than separately
  • Bioproduction teams selecting high-producer cell lines to shorten biopharmaceutical manufacturing timelines

When not to use Screening Hub

  • Engineering teams that need an API to plug screening analysis into an automated pipeline: none exists
  • Buyers who require a published price list or self-service sign-up before speaking to a sales contact
  • Clinicians looking for a diagnostic tool, since the platform addresses research and drug development only
  • Laboratories whose instruments do not export PerkinElmer Harmony files, the only input format documented
  • Procurement teams that need published terms, a data processing agreement or a security certification upfront
Get started

How to use Screening Hub

A typical end-to-end flow, from setup to results.

  1. Get in touch through the contact form, or work through the four-question quiz on the site to describe your stage, goal, data type and main challenge
  2. Receive an individual quotation: no price is published, so the scope and cost are set project by project
  3. Take part in the needs analysis, where research objectives and data collection strategies are defined jointly
  4. Hand over your raw screening data or experimental results; PerkinElmer Harmony exports are the documented input format
  5. Let the vendor standardise and optimise the image data before analysis
  6. Choose the relevant module: Classic for high-content screening, Transcriptome for combined omics and imaging, 3D for tissue sections, High Producer Cells for bioproduction
  7. Have the deep-learning models run the analysis and surface candidate compounds or cells
  8. Review the visualisations, including cell proportions across wells for the same treatment
  9. Read the interpretation of biologically relevant parameters such as structural changes and modes of action
  10. Receive the final analysis report with validated results, and use the demo plus optional onboarding to get started
Quick read

Pros & Cons

Pros

  • Four distinct modules covering 2D imaging, transcriptomics, 3D tissue data and bioproduction, a rare breadth
  • Named, checkable references: Institut Pasteur Paris, Fraunhofer IGB, aimed analytics, 420 Pharma, ZIM and Eureka programmes
  • Scientific positioning argued from a published study in Trends in Cell Biology rather than simply asserted
  • Full service with no implementation effort: the vendor runs the analysis end to end
  • No integration with existing interfaces required, which lowers the technical barrier considerably
  • A quantified accuracy claim, around ten percent better classification than conventional machine learning
  • German publisher, processing within the EU, with GDPR compliance claimed explicitly

Cons

  • No published price at all, and the vendor states its cost structure is still being finalised
  • No API and no documentation subdomain, so the analysis cannot be automated or embedded
  • No terms of service anywhere on the site, so no contractual commitment can be read before requesting a quote
  • No data processing agreement, no sub-processor list and no security certification of any kind
  • The hosting location of customer data is never disclosed
  • Only one input format is documented, PerkinElmer Harmony exports, and the High Producer Cells module is still under development
  • Signs of neglect: last substantial content update in April 2025, a single blog post from July 2024, a published but empty demo page and a broken menu link
Pricing

Pricing & Plans

No price is published. The vendor states that the scope of its services depends on the size and specific requirements of each project and that an individual quotation is provided on request, and its own FAQ confirms that the cost structure is still being finalised — usage-based billing, flat rates, monthly fees indexed on AWS and platform usage, per-user fees and surcharges for additional modules such as Explainable AI are all described as options under consideration. Two introductory offers are presented. OmicsVision is offered as an initial analysis at a fixed price, the amount of which is not disclosed. PhenoVision is described as freemium: the Smart Screening consulting package and the PhenoVision module are free, with a Smart Insights report package available at a fixed, undisclosed price. No free trial of the software is offered, and no currency or billing unit is stated anywhere.

OmicsVision — Smart Screening analysis of transcriptome and image data
  • the Smart Screening consulting package
  • the OmicsVision AI module and a results report
  • at a fixed price that is not published
PhenoVision — Smart Screening analysis of 2D and 3D cell images
  • a free Smart Screening consulting package with the PhenoVision AI module
  • plus an optional Smart Insights report package at a fixed price that is not published
Special offers — OmicsVision welcome offer: an initial Smart Screening analysis of transcriptome and image data at a fixed price, amount not published · PhenoVision welcome offer, presented as freemium: a free Smart Screening consulting package together with the PhenoVision AI module, with a Smart Insights report package available at a fixed price as an option
Prices and plans listed above may evolve. Always check the official pricing page before subscribing.
Trust & Privacy

Data, GDPR & hosting

A consolidated view of how Screening Hub handles your data.

GDPR overview

GDPR compliance is claimed explicitly, but in a single sentence on a single page. The FAQ on the get-started page states that the platform complies with GDPR requirements and follows a Secure-by-Design principle, supported by strong identity and access management, regular backups and tenant isolation. The privacy policy itself covers only the website: it lists data subject rights with their articles — access (15), rectification (16), erasure (17), restriction (18), portability (20), objection (21) and complaint to a supervisory authority (77) — and gives mail@valuedata.io as the contact for data protection matters. It carries neither an effective date nor a version number. No data protection officer is named, no data processing agreement is offered and no sub-processor list is published. As the publisher is established in Germany, no Article 27 representative is required.

Who owns the data?

The site publishes no terms of service and no data processing agreement, so ownership of the data submitted to the platform is never stated in writing. The privacy policy that does exist covers the website alone: it explains that details sent through the contact form are kept only for as long as the enquiry is being handled and are then deleted. At platform level the vendor's commitments are technical rather than contractual — a stated Secure-by-Design approach, strong identity and access management, regular backups and tenant isolation. Anyone sending proprietary screening data should settle ownership explicitly in the individual quotation, because nothing published on the site does it for them.

Reuse rights

Nothing on the site tells customers what they may do with the results afterwards, for the simple reason that no terms of service exist. The deliverable described is an analysis report of validated results together with visualisations, produced by the vendor from the customer's own image and transcriptomic data, so reuse is a matter for the individual contract rather than for any published licence. The vendor never states whether customer data is used to train its models: the subject is not raised anywhere, and no opt-out is offered. The only opt-out on the site is the Google Analytics browser add-on, which concerns website visitors and has nothing to do with platform data.

Data retention & training

Retention summary
No retention period is published for data held in the platform. The only retention rule on the site concerns the website contact form: submissions are kept for as long as the enquiry requires and are deleted once it has been handled. For cookies, the privacy policy explains that most are session cookies removed at the end of a visit, while others remain until the user deletes them. No figure in days, months or years appears anywhere, and the privacy policy carries no effective date. The right to erasure is restated under Article 17 GDPR, again without any stated deadline. Retention terms for screening data therefore have to be agreed in the individual contract.
GDPR contact

Hosting summary

The site does not say where customer data is hosted. No country, no region and no cloud provider is named for the platform. Amazon Web Services is mentioned once, and only as a possible basis for pricing — monthly fees indexed on AWS costs and platform usage — which is not a hosting statement and should not be read as one. The guarantees that are published are organisational rather than geographic: a Secure-by-Design principle, strong identity and access management, regular backups and tenant isolation. The publisher is established in Germany and German law governs the site, which makes EU processing the reasonable expectation, but it remains an expectation and not a declaration. The marketing site itself runs on IONOS infrastructure in Frankfurt, which says nothing about where the analysis platform runs. The only transfers outside the EU that are documented concern Google Fonts and Google Analytics on the website. Anyone with a residency requirement should ask for it in writing.

Watch-outs

Things to keep in mind

Risks and trade-offs to weigh before adopting Screening Hub.

  • No terms of service exist on the site, so nothing published defines who owns the screening data you submit or what may be done with it: settle this in the quotation before sending anything proprietary
  • The vendor never says whether customer data is used to train its models, and offers no opt-out, so a laboratory sending unpublished compound data is relying on trust rather than on a written commitment
  • The hosting location of customer data is never disclosed, and no data processing agreement or sub-processor list is available, which makes an internal privacy review difficult to complete
  • No security certification is claimed at all — neither ISO 27001, nor SOC 2, nor any GxP qualification — so the Secure-by-Design statement cannot be verified against an external audit
  • Pricing is not merely hidden but unfinished by the vendor's own account, so an early quotation may not reflect what the service will cost once the model is settled
  • Treating an automated hit list as a decision rather than a hypothesis is the main scientific risk: the vendor's own material presents the platform as a way to prioritise candidates, not to replace experimental validation
  • Several signals suggest the product is receiving little attention: the last substantial site update dates from April 2025, the demo page is published but empty, a navigation link returns an error, and the publisher's own corporate site no longer mentions the brand
Setup

Setup & Integrations

Technical difficulty

Technically, very little is asked of the customer. There is nothing to install, and the vendor states that no integration with existing interfaces is necessary; it takes on the whole analysis, from needs definition to reporting. Users upload raw screening data, with PerkinElmer Harmony exports as the documented input format, and new customers receive a demo plus optional onboarding. An online tutorial and manual are still in development. The real friction is commercial rather than technical: there is no self-service entry, no published price and no terms to read, so every engagement starts with a quotation.

Deployment

Web app
Company

Behind Screening Hub

Company name
ValueData GmbH
Founded
23/12/2019
Country of origin
🇩🇩 Germany
Headquarters
Overather Str. 62, 51429 Bergisch Gladbach – Bensberg, Germany
UBO
INFORMATION_NOT_FOUND
UBO country
INFORMATION_NOT_FOUND
Domain registrar country
INFORMATION_NOT_FOUND
Legal contact
Official links

Resources

All the official URLs gathered for verification and reference.

FAQ

Frequently asked questions

What does Screening Hub actually do?
It analyses microscopy images produced by cell screening campaigns. Deep-learning models read thousands of images automatically, classify the distribution of cell types and identify compounds that trigger a measurable cellular effect, so that promising candidates can be prioritised earlier in drug development.
Which modules are available?
Four. Classic runs high-content screening with an advanced Cell Painting approach. Transcriptome combines microscopy screening with transcriptomic data through multimodal clustering. 3D processes tissue sections using 3D convolutional and graph neural networks. High Producer Cells targets biopharmaceutical production and is described on the solution page as still under development.
How much does it cost?
No price is published anywhere on the site. The vendor states that the scope of service depends on the size and requirements of each project and that an individual quotation is provided on request. Its own FAQ adds that the cost structure is still being finalised, with usage-based billing, flat rates, per-user fees and add-on modules all under consideration.
Is there a free plan or a free trial?
Neither is offered as such. The PhenoVision welcome offer includes a free Smart Screening consulting package alongside the AI module, with an optional paid report package — but that is an introductory offer, not a permanent free tier, and the site never mentions a trial of the software itself.
Is there an API?
No. Every API path tested returns a genuine 404, and no api, docs, developers or app subdomain resolves. The site itself argues the opposite way round, presenting interface independence as a benefit: no integration with existing interfaces is necessary.
What data can I feed into the platform?
Users upload raw screening data or experimental results. The only input format the site documents is Harmony exports from PerkinElmer. The vendor says it is exploring options for transferring data back into existing systems for easier export of reports and results.
How fast is it, and how many images can it handle?
The vendor describes capacity as theoretically unlimited, depending on the number and type of GPUs used: sixteen images per run on GPUs with 16 GB of VRAM, thirty-two per run on 24 GB. Analysis time varies with the GPU configuration.
How accurate is it compared with traditional methods?
The vendor claims its deep-learning methods improve classification accuracy by about ten percent over conventional machine learning, citing a study titled Deep learning in image-based phenotypic drug discovery published in Trends in Cell Biology. Further validation studies are described as in progress.
What support and onboarding are included?
New users are given a demo and, optionally, an initial onboarding session. Technical support is available. An integrated online tutorial and a manual are described as still in development.
How is sensitive biological data protected?
The vendor states that the platform follows a Secure-by-Design principle and complies with GDPR requirements, with strong identity and access management, regular backups and tenant isolation. No data processing agreement, sub-processor list, hosting location or security certification is published, so the detail has to be requested.
Conclusion

Should you pick Screening Hub?

Screening Hub occupies a narrow and genuinely technical niche: turning high-content microscopy screens into ranked drug candidates, with four modules that between them cover 2D imaging, transcriptomics, 3D tissue data and cell-line selection for bioproduction. The named references — Institut Pasteur, Fraunhofer IGB, aimed analytics, publicly funded ZIM and Eureka projects — are unusually concrete for a tool of this size, and the scientific positioning is argued from a published study rather than asserted.

What is on offer is a service, not a product you can try. The vendor takes on the whole analysis, no integration with existing interfaces is required, and the only documented input is a PerkinElmer Harmony export. That suits a laboratory with a one-off screening campaign and no internal image-analysis capacity. It suits far less a team that wants to automate: there is no API, no documentation subdomain, no self-service.

The commercial and contractual side is where caution is due. No price is published anywhere, and the vendor says so itself — the cost structure is described as still being finalised. There are no terms of service on the site at all, no data processing agreement, no list of sub-processors and no security certification; the privacy policy is a standard German template covering the website alone, without an effective date. Where customer data is hosted is never stated.

Two further signals deserve weight before committing. The site's last substantial content update dates from April 2025 and its only blog post from July 2024, while the publisher's own corporate site, relaunched as a data science CRO, no longer mentions the brand at all. Worth a conversation — but ask for the contract, the hosting location and the current roadmap before sending any proprietary data.