Week 2 - McKellar

My week has been spent filling out paperwork, trying to get HSS to let me into clinic, and coding. Instead of describing that, I would rather spend a brief moment talking about a recent shift in my perspective on biology (inspired by a few chance events that occurred over the past two weeks).

On the bus to NYC, I opened up a section of Carlo Rovelli's "The Order of Time"- a wonderful book on our ever-changing understanding of what time is. I picked up where I had left off- in chapter 6, "The World Is Made of Events, Not Things". Here Dr. Rovelli describes his thoughts on why our view of nature should be framed by processes and events rather than collections of objects. Even stones are subject to the dynamics of time. Every seemingly permanent object is really just a series of thermodynamically favorable events, ultimately leading to change.

Just a few minutes prior to picking this book up again, I glanced through a new bioRxiv preprint from Lior Pachter's lab at Cal Tech (https://www.biorxiv.org/content/10.1101/658401v1). The paper describes their adaptation of a recent bioinformatics method (http://velocyto.org/) whereby they use single-cell RNA sequencing and surface protein expression data to predict the timing of gene expression (this work is really exciting to me and deserves more than a short summary; maybe I’ll get to that later this summer).

As someone who studies single-cell biology, the juxtaposition of these two pieces of work lead me to reflect on how I think about cells. So much of the work we have done has focused on identifying the contents of certain cells, how we can define each cell type, and how we can use these classifications to build biological references. So much of the literature has focused on descriptions of cells, rather than what is actually happening inside them. Single-cell biology (in its current form) is very young and this bias toward descriptive work is largely driven by technical limitations.

This past weekend I visited the Museum of Modern Art (MoMA, on the last day it will be open this summer, luckily), and another fortunate juxtaposition got me thinking. MoMA happens to have a collection of works by Georges-Pierre Seurat. My fellow single-cell enthusiasts might know him by a recent software package named after him (https://satijalab.org/seurat/). The seurat software package is used for analyzing single-cell RNA sequencing (and more recently, additional modalities) data. It’s named because of the figures it is commonly used to generate, where individual cells are projected onto a 2D plot, grouped together based on gene expression patterns, and visualized as small dots. An example from the Cosgrove Lab’s recent preprint is below:

Single-cell RNA sequencing of dissociated murine muscle tissue, generated by Andrea de Micheli (https://www.biorxiv.org/content/10.1101/671032v1) in the Cosgrove Lab at Cornell (http://cosgrovelab.bme.cornell.edu/).

Perhaps Rahul Satija named the package after seeing the same Seurat paintings in MoMA. It’s easy to see similarities in Seurat’s post-impressionist pointillism and today’s UMAP and tSNE plots.

Just around the corner, MoMA has a room filled with Monet paintings. Spanning an entire wall, the three panels of Water Lilies draw huge crowds (https://www.moma.org/collection/works/80220). It was here, while pondering the genius of Claude Monet, that I remembered Rovelli’s ideas about events versus things. Monet’s blurred brush strokes can’t be teased apart by the naked eye. Each color is layered on top of another, creating a continuum of color across the canvas.

If we take the Rovellian viewpoint (an adjective I will take the initiative to add to English), biology should not and cannot be seen as a collection of dots. Intercellular interactions are lost, spatial relationships are lost, and our understanding of the events that drive biology are veiled. As I said before, technical hurdles limit the number of cells we can sample and the depth of the information gathered, but work (like that of the Pachter lab mentioned before, and many others) is beginning to shed light on the dynamic web of interactions that is biology. Instead, we should begin to think of cells as existing more like the brush strokes of Monet than those of Seurat. They are entangled, coexisting in, on, and around each other.

This idea of continuity and decompartmentalization is not new. A quote from the great physicist Ernest Rutherford proves that- "All science is either physics or stamp collecting." He clearly worked in the days of rock-collecting geologists and species-counting naturalists, but the man had a point. A somewhat Rovellian point, at that. The single-cell revolution has brought biology to a fork. One side is further collecting massive cell atlases (a term I personally dislike, considering most single-cell RNA datasets lack the defining component of an atlas- spatial information). Like a reference genome, these atlases (still don’t like this term) are incredibly useful in streamlining classification and description of biological samples. But at a certain level, it is just collecting more cells. The other direction, as you may have guessed, is to look at the relationships between these cells- the events. Inspired by Monet and Rovelli, I think that is the road I’ll travel (whether it is less-traveled or not).