Natanael SPISAK

Genomic sequencing reveals genetic variation at many scales: divergence between species, heritable differences among individuals, and the mosaic makeup of somatic tissues. Our group focuses on recent and rare variants, which are informative about the sources of genetic variability and comprise much of what causes genetic disease and cancer.

We develop mathematical models and statistical inference methods to learn about the biology of mutation and genetic disease from sequencing data.

Toward a comprehensive model of mutation

Mutations are the original source of genetic variation, yet we know surprisingly little about how they arise in vivo and what determines their rate and spectrum across different cell types.

The sequencing of large tumor cohorts has revealed repeatable patterns, termed mutational signatures, and helped link some of them to specific mutagenic processes. However, the implicit model treats each process as producing a single, invariant signature, largely ignoring that signatures reflect a net result of DNA damage and repair. When repair is compromised, as in individuals with constitutional repair defects or in certain tumors, the underlying damage is exposed more directly. Such cases carry vital information that current approaches do not fully capture. We aim to develop a model that will allow us to learn from these cases systematically.

In a complementary direction, we collaborate with Thomas Dupic (Centre d'Immunologie de Marseille-Luminy) on somatic hypermutation in B cells. In this process, targeted damage and error-prone repair locally elevate the mutation rate by a few orders of magnitude. The fast pace and well-characterized source of damage make it a natural model system to study the kinetics of mutagenesis.

Somatic variation and the genetics of disease

A large fraction of suspected genetic diseases cannot be linked to a causal variant. What are we missing? The answer likely involves regulatory variation, which remains difficult to interpret, and somatic mutations present in a fraction of cells only. We are interested in estimating how often somatic variants play a role in genetic disorders. De novo germline mutations known to contribute to disease provide a point of reference: what do the variants with known effects imply about those we have yet to identify? On the data side, we work on identifying somatic mutations from long-read sequencing, where single-molecule resolution makes it possible to call somatic variants reliably at scale.

Environment

We work alongside the groups of Antonio Rausell and Matthew Ricci, whose complementary expertise lies in statistical modeling and machine learning, as well as Sara Bizzotto's laboratory, investigating somatic mutations in neurodevelopment. We also draw on the institute's genomics and bioinformatics platforms.

Contact

If these research questions appeal to you, or you have others to bring, we would be glad to hear from you. We have open positions for students and postdocs.