Thanks for sticking around to the end (here’s a bonus image as a treat)! Would love to hear what you think! 16/16

We also found that SFRP1 had a strong clonal gene expression pattern, with much higher expression in one particular clone compared to the rest of the tumor or even its neighbors. 13/16

With this, we discovered several genes with spatially-driven expression patterns. For example, MITF was expressed on the exterior of the tumor and VEGFA was expressed on the interior, near the vascularized region. 12/16

This allowed us to visualize the pattern of clones in the tumor! 10/16

Here's an example of SpaceBar in action! Each cell has transcripts from multiple barcodes (depicted with dots) and we can identify sister cells by their barcode combos. 7/16

For SpaceBar, we designed a library of 96 barcode sequences and inserted them into cells such that individual cells received multiple copies. Then, by reading out the combination of barcodes, we can identify cells that are descended from the same initially labeled cell. 6/16

Excited to share SpaceBar - our new method for labeling and detecting clones with imaging-based spatial transcriptomics platforms! w/ Yael Heyman and @arjunraj.bsky.social www.biorxiv.org/content/10.1... 🧵

Some of these lineages do appear to be heading down a road towards specialization.

However, others have improved performance in both growth phases, albeit via a series of steps that tend to improve one or the other, suggesting generalists can still evolve. 9/n

While 85% of 1st-step mutations improve both fermentation and respiration performances, only 35% of second-step mutations do so.

This suggests that already in the 2nd-step of evolution, we have begun to run out of mutations that improve both growth phases to a large extent. 6/n

To do this, we conducted a series of 2nd-step evolution experiments.

We constructed barcoded populations from five 1st-step mutations and evolved them in the same environment as the 1st step.

We then isolated 100s of mutants and quantified their performances in growth phases. 5/n

We wondered whether this pattern might continue over further adaptive steps, suggesting that mutations that improve multiple traits at once are common (left).

Or if we begin to run out of these mutations, mostly finding mutations that improve one trait or the other (right). 4/n

For instance, Yuping previously found that ~85% of first-step adaptive mutations improved yeast's performance in both fermentation and respiration growth phases under selection in this environment. 3/n