A lineage-based model of scalable positional information in vertebrate brain development
How the cells do know where they need to be and what they should be, where to go (which cell to become) and with whom to be linked in the brain?
Probably, you would answer like me "it's done by signaling", and we are kinda right, right? but this new research claims it could be done by the cells due to their gen
etic memory, where daughter cells inherit gene expression from their progenitor cells, rather than by local cellular signaling.
Development from a zygote to an adult organism involves complex interactions among thousands of genes. These genes exhibit highly dynamic expression across space and time. Here we report a striking simplicity amidst this complexity: Despite individual gene expression variability, the eigengene—the first principal component of gene expression—exhibits an invariant global spatial pattern throughout the embryonic and post-natal stages of the mouse brain.
Furthermore, the mouse pattern is observed also in the larval zebrafish, revealing that eigengene expression is conserved over 400 million years of evolution. We show that the eigengene pattern can be explained by a simple lineage model in which daughter cells’ gene expression is similar to that of their parent, but cannot be explained by one in which gene expression arises through local cellular signaling.
A few moments I want to focus on a bit more:
-
Clustering results show that the temporal state of the brain is the dominant factor in single-gene expression profiles, causing voxels (volumetric pixels) to group with others from the same age.
-
Diffusion is only effective over short distances (50 - 100 cells), making it difficult for local interactions to create persistent global gradients without complex relays.
The pipeline is:
1️⃣ Data Collection -> 2️⃣ Filtering & Normalization -> 3️⃣ PCA (Principal Component Analysis) -> 4️⃣ Spatial Mapping -> 5️⃣ Cross-Projection & Hierarchy
1️⃣ They gathered 3D spatial gene expression data (measured in "voxels") from mouse brains across multiple embryonic and post-natal stages, as well as from larval zebrafish brains
2️⃣ For the mouse data, they filtered out genes with too many missing measurements, leaving a core set of 1256 genes.
3️⃣ They organized the data into a matrix of voxels versus genes and applied PCA at each developmental stage. The first principal component (PC1), which they termed the "principal eigengene"
4️⃣ They mathematically projected this eigengene back onto the physical brain voxels.
This revealed how strongly the "master pattern" was expressed in different physical locations
5️⃣ To test for stability, they projected the eigengene calculated from an early stage (E11.5) onto the data of later stages and onto the zebrafish data.
—-
And what do you think about this way of looking at brain development mechanics?