Research Summary Plant Science Meets Mathematicians
As a mathematician, I support the Danforth Center mission to improve the human condition through plant science, but my contribution is unique from most other scientists at the center. Although from different disciplines, we mathematicians, plant biologists, computational scientists, and statisticians work together to achieve a common goal, and try to address problems in plant science such as environmental sustainability, pest and disease resistance, enhancing drought tolerance, and increasing yield in a rapidly growing population. These are big problems and require the interaction of researchers, new technology, and data analysts. When plant science meet mathematicians, it means collaboration! Data Meet Mathematicians
In recent decades, imaging technologies have developed incredibly quickly. These remarkable advances allow us to collect vast amounts of high context data every day, awaiting measurement and analysis. But can we simply use our eyes to look at countless pictures and accurately judge, for example, which plants are bigger or smaller, or more importantly how different they are? We might use a ruler to manually measure features from the images, such as the length of every branch. However, you can imagine how labor intensive this is! In addition, there are many plant structures such as sorghum panicles that are so complex that most of the features are not tractable with manual measurements. Therefore, we need to employ mathematics and computational tools to process images, extract features, and analyze data. We need to do it efficiently and precisely, to build models to predict the future, and to fully utilize new technologies. This will accelerate discoveries and our understanding of how to modify the plant form to improve agriculture. When data meet mathematicians, it means precision, prediction, and prescription! Mathematicians meet Data
Data are collected at different scales. For example, they can be microscopic images showing cells, or can be field pictures taken by drones. Data are also collected from different organs - for instance, they can be 2D scans of leaves or can be 3D images of roots. Finally, data can be collected dynamically, such as with a series of images showing plant growth or cell expansion. Data always puts a glint in my mathematician’s eyes, because we see them differently and comprehensively. When mathematicians meet data, it means something new will be developing! Quick glimpse
- 2D (leaf) shape: a persistent homology based mathematical approach was developed that can be applied to any 2D shape such as leaf shape. It summarizes more leaf shape information into a single descriptor, capturing a large amount of otherwise undetected features, variation, and genetic associations.
- 3D inflorescence/root/branching architecture: an advanced framework to comprehensively quantify 3D branched plant structures was developed, allowing computational dissection of plant organs. This enables capabilities such as the identification and isolation of individual seeds from sorghum panicle, the extraction of geometric and topological (e.g. persistent homology) based features, and the building of structural simulations such as grape berry potential.
- 3D genome structure: as part of an ongoing project, we will develop a new method to characterize 3D genome structure and study how plant genome structure changes under different CO2 levels.