Seed Secrets Revealed: Cracking the Microbiome Code

Scientists have discovered a way to predict the types of bacteria that live on seeds, based on the plant's family tree. This breakthrough uses machine learning to forecast the seed microbiome, which is the community of microorganisms that live on the seed. By analyzing the plant's DNA, researchers can now guess which bacteria are likely to be present on the seed.

They tested this approach on 61 plant species and found that it works pretty well, especially for plants that are closely related. The method uses a special type of distance measurement to compare the plants' DNA and predict which bacteria will be present.

The researchers found that one particular model, called HD-KNN, was really good at making these predictions. It was especially good for plants that have a lot of close relatives in the dataset. For plants that are a bit more isolated, another model called HD-GPR worked better.

This new approach could be really useful for understanding the microbiomes of plants that haven't been studied yet. It could help scientists prioritize which plants to study next and make predictions about their microbiomes.

But it's essential to remember that this only works for seed-associated bacteria, and not for other types of microorganisms or plant parts. The environment plays a significant role in shaping these communities, and different parts of the plant have different types of microorganisms living on them.