Cristiana Banila
Co-founder & CSO
Mitrabio
Biography
Cristiana is the CSO and Co-founder of Mitra Bio, a Khosla Ventures backed start-up focused on non-invasive skin diagnosis for precision treatments using AI and epigenetics data. Within Mitra Bio, she leads the translation of the technology into industrial settings, currently working in partnership with top skincare and pharma companies to optimize product development and claim substantiation. Cristiana graduated from Oxford and Princeton Universities where she studied Biochemistry and Public Health Economics. She specialized in molecular diagnostics development, having earned a PhD from Queen Mary University of London on developing a self-sampling epigenetics-based test for early cancer detection. The test is currently in clinical trials for implementation in the NHS cervical screening programmed. Cristiana has been named Forbes 30 under 30 Europe, Science and Healthcare in 2023.
Conference
Day 1
Session 2: Exposome impact on Skin & Hair
Skin epigenetic clock to measure skin age and environmental exposures
Background: Ageing and environmental exposures are associated with reproducible alterations in DNA methylation (DNAm). Over the past decade, DNAm has underpinned the development of epigenetic clocks of ageing and exposure biomarkers that capture signatures of smoking, air pollution, diet, and inflammation across internal tissues and accessible fluids (e.g., blood, saliva). By contrast, DNAm research in skin has been relatively limited and has primarily relied on invasive biopsies and platforms not optimised for real-world skin assessment. Tape stripping provides a non-invasive alternative for skin DNAm research, enabling robust collection of epidermal material at scale without the need for specialised clinical personnel.
Methods: We profiled epidermal DNAm using enzymatic methyl-sequencing (EM-seq) from non-invasive facial and body tape strips, assembling the largest non-invasive human epidermis dataset (n = 462). We collected extensive lifestyle and demographic data including chronological age, sun exposure, and smoking habits. Two epigenetic age models were trained with PCA-regularised Elastic Net regression: MitraSolo (single CpGs) and MitraCluster (co-correlated CpG regions). Performance was evaluated on independent test sets, longitudinal repeats, and external datasets. We benchmarked against five established clocks, e.g. Horvath Skin & Blood, and examined DNAm features related to exposome factors (e.g., anatomic site, photoexposure).
Results: Non-invasive tape-strip EM-seq yielded high-quality epidermal methylomes suitable for robust modelling. Both MitraSolo and MitraCluster estimated biological age directly from facial skin with accuracy that rivaling biopsy-based methods, and matching / outperforming established clocks across multiple validations. DNAm variation captured signals consistent with intrinsic ageing and environmental influence, supporting the suitability of tape-strip methylomes to interrogate the skin exposome alongside ageing.
Conclusions: Non-invasive tape-strip EM-seq enables scalable, longitudinal measurement of epidermal DNAm and accurate estimation of skin biological age directly on the skin. This platform supports exposome-aware assessment in trials and population studies, enabling molecular claim substantiation without biopsies.