Our engagements

Our engagements

From framing to training your teams

Discuss an engagement

What stays the same from one engagement to the next

The three engagements below share neither sector, nor technique, nor duration. They do follow the same way of working.

We start from your problem

A framing phase with your teams always precedes development: which question are we trying to inform, and what can answer it.

We validate on your real data

No demonstration on a hand-picked dataset. Results are tested against your production data before delivery.

Our own experts do the work

PhDs and omics specialists work directly with your researchers and production teams, not from behind a service desk.

We leave you autonomous

The engagement ends not at delivery but when your teams are trained on the tools delivered — which remain yours.

Biomedical research · Genetic diseases

Institut Imagine, AP-HP

Contextualising a cellular digital twin from single-cell data

Context

Institut Imagine works on genetic diseases. Its teams produce single-cell transcriptomics data (scRNA-Seq) of great richness, but analysing it remained largely a manual task.

The challenge

Connecting those measurements to a model of cellular metabolism — moving from a list of expressed genes to a reading of which metabolic functions are actually active in the cell.

How we worked

  1. 1

    Understanding the scientific question

    Before a single line of code, framing work carried out with the research team: which biological hypothesis are we trying to inform, and what in the data can answer it.

  2. 2

    Analysing the omics data

    Processing the scRNA-Seq datasets and preparing the data for model contextualisation — the least visible part, and the one everything else depends on.

  3. 3

    Developing the algorithms

    Writing the algorithms that contextualise the cellular digital twin: making a generic model carry the particular state of the cells measured.

  4. 4

    Studying metabolic functions

    Using the contextualised twin to study cellular metabolism, and to guide which experiments are worth running at the bench.

What the client keeps

  • A cellular digital twin contextualised on the team’s own data
  • The contextualisation algorithms, documented
  • A reading of the active metabolic functions, usable upstream of discovery

Team on the engagement

A PhD in AI and omics, and an MSc in omics, working directly alongside the researchers.

« A formidable team to work with, specialists in metabolic digital twins, they offer a collaborative, intuitive and creative solution. »
Mickael MENAGER — Institut Imagine, AP-HP
  • Digital twin
  • Single-cell transcriptomics
  • Metabolomics
  • Genetic diseases

Agronomy · Testing laboratory

Conidia-Coniphy

FungiResist™: from nanopore sequencing to the treatment decision

Context

Conidia-Coniphy analyses fungal resistance to fungicides. Nanopore sequencing gives them access to the variants present in a plot; turning that into a recommendation a grower can act on is another matter.

The challenge

Building the tool that turns sequencing data into a treatment decision — targeting only the resistances actually detected, and spotting rare variants before they spread.

How we worked

  1. 1

    Diagnostic

    A survey of the data produced, the processing in place, and what the teams actually expected from the tool.

  2. 2

    Advice

    Defining the solution with the client: what it had to compute, whom it addressed, and what it would not do.

  3. 3

    Implementing the algorithms

    Building the resistance analysis pipeline, from reading the sequences through to prediction.

  4. 4

    Analysis and validation on real data

    Testing the results against the laboratory’s own real datasets — a validation step, not a demonstration.

  5. 5

    Delivery

    Bringing the FungiResist™ application into service, operated by Conidia-Coniphy under their own brand.

  6. 6

    Training the teams

    Training the client’s teams on the delivered tools. An engagement ends not at delivery but when the people using it are autonomous.

What the client keeps

  • The FungiResist™ application, in service and under the client’s brand
  • The resistance analysis algorithms, validated on real data
  • Teams trained and autonomous on the tool
« We committed to a long-term collaboration with Gencovery and the support on this high-value project is excellent. Gencovery brings a strong, innovative vision on data management. »
Sébastien VACHER — CEO, Conidia-Coniphy
  • Nanopore sequencing
  • Genomics
  • Epidemiology
  • Predictive models
FungiResist™ on the Conidia-Coniphy website

Bioproduction · Fermentation

Greencell

Optimising a fermentation process, from project funding to digital twins

Context

Greencell develops fermentation processes to produce biomolecules. Every run produces hundreds of files — analyses, culture logs, measurements — scattered across xls, csv and txt.

The challenge

Moving beyond file-by-file analysis to reason about the process itself: understanding what drives yield variation, and knowing which experiments are worth running.

How we worked

  1. 1

    Setting up the collaborative project

    Building the application and securing funding alongside the client: a collaborative project with €800k of overall funding, backed by Bpifrance and the Auvergne-Rhône-Alpes Region.

  2. 2

    Experimental design

    Designing the experimental plan with the production teams, so that the data produced answers the questions asked rather than simply accumulating.

  3. 3

    Automating collection

    Connecting to the existing instruments and files, then structuring and annotating the fermentation data automatically.

  4. 4

    Building the models

    Building digital twins of the process, to simulate culture media and production conditions before testing them.

  5. 5

    Data analysis

    Reading kinetics, yields and production deviations, handed back to the teams in a form they can interrogate themselves.

What the client keeps

  • A funded, structured collaborative project
  • Automated data collection, from instrument to dashboard
  • Digital twins of the fermentation process
  • Documented, versioned and reusable process models
« A partnership to optimise bio-processes. »
Assia DREUX-ZIGHA — Director, Greencell
  • Digital twin
  • Bioprocessing
  • Experimental design
  • Machine learning
A platform built for life sciences

Ready to structure your projects differently?

See how Constellab can simplify your projects, strengthen traceability and unlock the full potential of your scientific data.

Request a demo Talk to an advisor
Free No commitment Quick cancellation