Predictive Modeling of Pseudomonas syringae Virulence on Bean using Gradient Boosted Decision Trees

Abstract

Pseudomonas syringae is a genetically diverse bacterial species complex responsible for numerous agronomically important crop diseases. Individual P. syringae isolates are typically given pathovar designations based on their host of isolation and the associated disease symptoms, and these pathovar designations are often assumed to reflect host specificity although this assumption has rarely been rigorously tested. Here we developed a rapid seed infection assay to measure the virulence of 121 diverse P. syringae isolates on common bean (Phaseolus vulgaris). This collection includes P. syringae phylogroup 2 (PG2) bean isolates (pathovar syringae) that cause bacterial spot disease and P. syringae phylogroup 3 (PG3) bean isolates (pathovar phaseolicola) that cause the more serious halo blight disease. We found that bean isolates in general were significantly more virulent on bean than non-bean isolates and observed no significant virulence difference between the PG2 and PG3 bean isolates. However, when we compared virulence within PGs we found that PG3 bean isolates were significantly more virulent than PG3 non-bean isolates, while there was no significant difference in virulence between PG2 bean and non-bean isolates. These results indicate that PG3 strains have a higher level of host specificity than PG2 strains. We then employed machine learning to investigate if we could use genomic data to predict virulence on bean. We used gradient boosted decision trees to model the virulence using whole genome kmers, type III secreted effector kmers, and the presence/absence of type III effectors and phytotoxins. Our model performed best using whole genome data and was able to predict virulence with high accuracy (mean absolute error = 0.05). Finally, we functionally validated the model by predicting virulence for 16 strains and found that 15 (94%) had virulence levels within the bounds of estimated predictions. This study demonstrates the power of machine learning for predicting host specific adaptation and strengthens the hypothesis that P. syringae PG2 strains have evolved a different lifestyle than other P. syringae strains. AUTHOR SUMMARYPseudomonas syringae is a genetically diverse Gammaproteobacterial species complex responsible for numerous agronomically important crop diseases. Strains in the P. syringae species complex are frequently categorized into pathovars depending on pathogenic characteristics such as host of isolation and disease symptoms. Common bean pathogens from P. syringae are known to cause two major diseases: the halo blight disease, which is characterized by large necrotic lesions surrounded by a chlorotic zone or halo of yellow tissue; and the bacterial spot disease, which is characterized by brown leaf spots. While halo blight can cause serious crop losses, bacterial spot disease is generally of minor agronomic concern. The application of statistical genetic and machine learning approaches to genomic data has greatly increased our power to identify genes underlying traits of interest, such as host specificity. Machine learning models can be used to predict outcomes from new samples or to identify the genetic feature(s) that carry the most importance when predicting a particular phenotype. Here, we implemented a rapid method for screening a proxy of virulence for P. syringae isolates on common bean, and used this screen to assess virulence of P. syringae strains on bean. We found that halo blight pathogens display a stronger degree of host specificity compared to brown spot pathogens, and that genomic kmers and virulence factors can be used to predict the virulence of P. syringae isolates on bean using machine learning models.

Plant phenotyping relevance

豆での病原性という植物の病態を測定する迅速な感染アッセイを開発し、その測定結果を用いた機械学習予測と機能検証を行っており、表現型取得法が中心的です。

Code and data availability

The supplied blocks describe the seed infection virulence assay, germination screen, and XGBoost modeling, but contain no public phenotype dataset deposit, no author code/scripts URL, no trained model release, and no supplement link. Mentions of Table S1 and custom Python scripts lack any availability statement or URL.

No evidence-backed public reproduction asset is currently recorded.