A tumor dataset does not answer a clinical question on its own. Which patients respond to checkpoint blockade, what separates their microenvironment from everyone else’s, and whether that signature survives the next cohort: this is the work GenXMap’s computational team takes on.

The questions we take on

Which patients respond to checkpoint blockade, and what separates them molecularly
How the immune composition of the microenvironment shifts under treatment
Whether a candidate signature holds on an independent cohort
How molecular findings line up with treatment arm and clinical outcome

How the analysis runs

Immune deconvolution

Cell-type composition estimated from bulk RNA-seq, cross-checked against single-cell references where the study provides them.

Single-cell annotation

Clustering, immune subtype annotation and differential abundance between conditions.

Cohort integration

Batch-corrected integration, so a cohort accumulates instead of restarting, with pipelines versioned per study.

Signature discovery and validation

Responder versus non-responder signatures, cross-checked on independent datasets before they leave the building.

Modelling

Models linking immune signature to outcome, with parameters, thresholds and versions recorded per study.

Working with a clinical cohort

Molecular results are reported against your clinical and treatment metadata, not as a standalone omics dataset.

Cohorts that enrol over time are integrated batch by batch into the existing analysis rather than reprocessed from scratch. Data handling is scoped to your consent and IRB framework at the start of the project, before anything is transferred.

What the report contains

Methods with exact tool versions, parameters and thresholds. QC and cohort-level figures. Statistical results. An interpretation written against the biological question you started from.

Built to be audited or reproduced by a third party.

Where the data comes from

The laboratory work behind these analyses runs on OMICS4, GenXMap’s platform site. Sample requirements, shipping and turnaround are published there.

We also take on data we did not generate. An existing dataset, from an earlier study or from another provider, is analysed on its own terms. What decides the work is the question and the metadata that comes with it, not who ran the sequencer.

Book the platforms on OMICS4

Which cells moved, and where they were

In immuno-oncology, knowing that expression changed is rarely enough. You need to know which immune subsets moved, and where they sit relative to the tumour. That is three modalities, not one: bulk RNA-seq on our own platforms, single-cell with Parean Biotechnologies in Saint-Malo, and spatial transcriptomics with Explicyte, a precision oncology and immuno-oncology CRO in Bordeaux. Run as one programme, on one design and one sample set, they are read together rather than reconciled afterwards.

GenXMap is also a member of the Paris-Saclay Cancer Cluster, with an office on site, and is joining the Marseille Immunology Biocluster. For a programme that needs a specific cohort or a clinical collaborator, that is the difference between a conversation and a search.

See the 360° transcriptomics offering

Immuno-oncology