Datharsis

Datharsis (Explain Tech S.L.) is a deep tech spin-off of the University of Granada, Spain, specialised in the explainable analysis of omics data across the life sciences, from health to agri-food and ecology. We help research teams design their studies and turn high-dimensional, noisy and multimodal data into traceable, reproducible evidence.

Scientific foundation. Our methods come from twenty years of peer-reviewed research at the Computational Data Science Laboratory (CoDaS Lab) of the University of Granada, and we use them under an exclusive commercial licence from the university. Our Chief Scientific Advisor, Prof. José Camacho, is a Full Professor at the University of Granada, head of CoDaS Lab and ranked in Stanford's top 2% of scientists worldwide. He leads our scientific strategy and our participation in European research projects.

Two things we do:

• Statistical design of experiments. Many studies fail not in the analysis but in the design. Before any sample is collected, we help define the questions the study must answer, the factors and interactions to include, the sampling strategy and the sample size needed, using our own computational methods. The result is more statistical power with fewer samples, and lower field, laboratory and sequencing costs. In a Horizon Europe proposal, this strengthens the methodology section from the start.

• Interpretable analysis and integration of omics data. We integrate omics layers (metabolomics, proteomics, transcriptomics, epigenomics, microbiota, spatial omics) with clinical, spectral or environmental data. Our models are interpretable by design, not explained afterwards with post-hoc layers. Methods developed by our team, such as vASCA, group-wise models and MEDA, reveal which variables drive each effect in experiments with several crossed factors. Cross-validation and permutation testing then confirm that the results are robust and not artefacts. See the scientific evidence.

Our team's methods have been applied to:

• Response to chemotherapy and survival signatures in cancer, using metabolomics

• Multi-omics profiling of a paediatric cohort, combining microbiota, metabolomics and clinical data

• The spread of a mycotoxin in wheat varieties over the crop cycle

• Food authentication and traceability

• The effect of mistletoe parasitism on the pine metabolome, separated from season, tree and canopy height

• The effect of contaminants across three tissues of crabs in ecotoxicology studies

• Warming trends in high mountain lakes, separating the shared climate signal from the features of each lake

These are only some examples. The same methods work wherever data combine several factors, many variables and few samples, and our team has also applied them to volcanic eruption forecasting, industrial process monitoring and network security. See all documented cases.

What we bring to a consortium:

• Leadership of the data analysis work package, or of specific analysis tasks

• Study design and sample size planning, from the proposal stage

• Discovery of interpretable biomarkers, indicators and signatures, with rigorous statistical validation

• Reproducible, auditable analysis pipelines aligned with FAIR principles

• DathaSuite, our analysis platform, so consortium researchers can explore their own data with the same methods

• Training of consortium researchers through our DathaXplain programme

If your project will generate complex omics data and needs results that hold up before reviewers and regulators, let's talk.