Michelle Chan wants to know: how does a cell change its fate? During animal embryogenesis, a small population of stem cells gives rise to all the cells and tissues of the adult animal. The progeny of these stem cells become progressively more differentiated and specialized as development progresses. But once a cell starts down a particular developmental path, is it always destined to trod a series of defined steps to reach a fixed destination, or can it take detours along the way? Might cells arrive at a common phenotype after traversing entirely different developmental paths?

Chan, who started her appointment with the Princeton Department of Molecular Biology and the Lewis-Sigler Institute for Integrative Genomics in September of 2020, trained as a computational biologist during her graduate studies in Aviv Regev’s lab at the Broad Institute at MIT and Harvard. There, she became interested in a project investigating the role of DNA methylation in mammalian embryogenesis.

“I worked with the very talented Zachary Smith, who’s now a PI at Yale, to generate the first genome scale DNA methylation maps for mouse and human early embryogenesis,” says Chan. “My graduate work was pretty much completely in computational biology, so for my postdoctoral work I joined Jonathan Weissman’s lab at UCSF in order to gain experience at the bench. One of the things that Jonathan’s lab specializes in is making new technologies, and that's what I did. I made a new lineage tracing technology that we call a ‘molecular recorder’.”

The molecular recorder that Chan designed works by first integrating a short stretch of “target” DNA into a mouse egg cell’s genome at multiple different sites, then generating small deletions or insertions of DNA, called indels, at these sites. As the fertilized egg divides into new and progressively more differentiated cell types, the new cells acquire additional indels. Indels that occur earlier in development are found in more cells later on, and cells that are closely related share more of the same indels. The lineage of different cells in the embryo can therefore be determined by comparing the indels they possess.

“We designed this system so indels can be traced with single-cell RNA-seq, which gives us a simultaneous readout of both the cell’s lineage and its molecular phenotype,” says Chan.                                                     

A cell’s molecular phenotype is determined by sequencing its transcriptome, which is the set of genes being expressed by the cell. The initial study, which sequenced the indels and transcriptomes of 10,000 cells from 8.5 day-old mouse embryos, yielded a surprising result.

“We identified an alternative path towards embryonic endoderm, which was surprising because the alternate origin is from a differentiation event that occurs very early in development,” Chan notes.

In her own lab at Princeton, Chan plans to apply approaches derived from both computational and experimental biology, including the molecular recorder, to the study of mammalian development. Students and postdocs in Chan’s lab will also have the opportunity to work on refining and improving the molecular recorder and related technologies, getting in on the ground floor of an exciting new phase in the field of developmental biology.

“The question I want to ask is, how do cells change cell fate? In mammalian development, it's still not known how you get to many of the cell types from pluripotency. One way to get to this question is to first find a cell fate map. By applying my lineage tracing technology to more embryos, we hope to uncover more of these differentiation pathways,” she says. “Once the cell fate map is identified, I’d like to study how robustness is encoded in mammalian embryogenesis. Mammalian embryogenesis actually has a lot of stochastic points because the process has to allow for plasticity in order to account for issues like environmental perturbations.”

In addition to getting her lab up and running, Chan is currently working with University colleagues Britt Adamson and Yuri Pritykin to design a new genomics course, to start next year, which will cover both computational and experimental aspects of the field.

“A lot of biology is becoming more quantitative in nature because a lot of technologies we use these days are capable of producing these huge data sets,” says Chan. “For future scientists, it is really important to get an understanding of computational biology, even if they don't necessarily do it themselves.”