Spatial biology labs can now capture many markers in a single tissue image. Turning those images into usable results is still slow, technical and expensive for most teams. In this Spotlight On, we look at Mantalys, a French start-up founded in 2025 that builds image analysis software for spatial biology.
Mantalys is working on a practical question: how can a lab analyse complex, high-plex tissue images without a cloud subscription, specialist hardware or a team of bioinformaticians?
Why spatial image analysis is still hard for most labs
Multiplex and high-plex immunofluorescence show cell types, tissue structure and how cells sit in relation to each other. But whole-slide images are large and difficult to process. For many labs, getting from an image to a result still means:
Specialist image analysis expertise.
Several separate software tools.
Cloud computing or GPU hardware.
Slow manual steps.
That makes it hard for non-specialists to produce results they can repeat and rely on. AI can speed things up, but scientists still need to see what the software is doing and correct it where needed. Mantalys' starting point is that good image analysis shouldn't require every researcher to become an image informatics specialist.
About Mantalys
Mantalys is based in Toulouse, France, and was founded in July 2025. Its founders are scientists with backgrounds in oncology, immunology, computer vision, robotics and AI.
The company's aim is to make high-performance image analysis tools that labs can afford, use easily and run without heavy computing resources. Its main product is MantaPlex, software for spatial analysis of multiplex and high-plex immunofluorescence tissue images. MantaPlex uses a hybrid AI approach that combines evolutionary computing with deep learning.
How MantaPlex works
MantaPlex runs locally on a Windows 10 or 11 computer. It needs no GPU, no cloud subscription and no moving data between tools. For labs handling large whole-slide images, that removes a real cost and a lot of admin. The software takes users through one workflow, in order:
Visualise the tissue image.
Segment it into individual cells.
Classify the cells.
Map where they sit and how they relate.
Detect structures and patterns.
Export the results.
Every step is recorded, so analyses can be repeated and checked. MantaPlex works with low-plex, multiplex and high-plex panels, and is designed for both experienced and new users. It is for research use only.
Key features
Automated steps throughout: The hybrid AI handles tissue detection, artefact removal, thresholding, clustering and structure generation, cutting time at each stage.
Researchers stay in control: Masks can be edited, segmentation settings adjusted and results corrected by hand. Users can see and change what the software has done.
Single-cell analysis: Several segmentation, thresholding and clustering options, with UMAP plots and single-cell feature tables for export.
Spatial interaction mapping: Proximity profiling, proximity graphs and nearest-neighbour analysis show how cells are arranged and which ones sit next to each other, as well as which cells are present.
Automatic structure generation: The software draws single- or multi-class structures around cell clusters, with adjustable settings and first- to third-degree neighbour identification.
Light on computing power: It runs on Windows 10 or 11 with no GPU. Using fewer computing resources was part of the design from the start.
Works with common imaging platforms: MantaPlex supports OME-TIFF, QPTIFF, Leica SCN, Zeiss CZI and OME-zarr formats.
Custom image analysis services
Mantalys also builds bespoke image analysis for labs and equipment manufacturers. Projects cover nanoparticles, organoids, tissue, whole-animal imaging, electron microscopy, confocal and intravital microscopy.
Using the same hybrid AI, the team can train a custom solution on as few as 1 to 10 annotated images. Results are reproducible and can be delivered as a no-code interface or as source code ready to integrate.
For teams that need to see exactly how results are produced, Mantalys can build fully interpretable pipelines using evolutionary computing instead of deep learning.
Why it matters for spatial biology
Spatial biology is moving out of specialist centres and into wider research use. Many of those new users won't have image analysis experts or large computing budgets.
Software that runs on a standard PC, shows its working and lets scientists correct it could help far more labs get answers from their spatial data. Labs can already produce far more spatial data than they can easily analyse, and tools like this help close that gap.
What it means for hiring
Young tools companies like Mantalys sit between biology, AI and software product development. As they move from launch to growth, the senior hires they make shape how fast they scale. Key roles at this stage include:
Commercial leaders who can build a sales channel for research software and reach labs, core facilities and pharma.
Product and technology leaders who can turn scientific methods into software that non-specialists can use.
Application science leaders who can support customers and feed their needs back into the product.
Experienced board members and executives who have taken a life science software company through early growth.
The same mix of scientific and computational skills is also needed in pharma and biotech teams adopting spatial biology in-house.
What's next for Mantalys
With MantaPlex now launching, Mantalys is bringing its software to researchers working with multiplex and high-plex tissue images. The team is building around what early users need, with work planned on:
Tools for handling arrays and batch processing.
New cell segmentation methods for more accurate analysis of irregularly shaped cells.
Ongoing workflow changes based on early user feedback.
You can find out more about MantaPlex on the Mantalys website.
—
