Predicting Symptom Severity in Infected Tomato Plants Using the Genome Sequence

Greg Howard
4th July, 2024

Predicting Symptom Severity in Infected Tomato Plants Using the Genome Sequence

Image Source: Natural Science News, 2024

Key Findings

  • Researchers in Japan developed an algorithm to predict the severity of viroid-induced symptoms in plants using genome sequences
  • The algorithm mimics the plant's natural defense process and clusters viroids based on how well they match the host genome
  • Validation experiments confirmed the algorithm's effectiveness, showing a strong correlation between viroid and host plant genome sequences
Viroids, the smallest known infectious agents, can cause a wide range of symptoms in plants, from latent infections to severe disease. This variability in symptom severity poses a significant challenge to agriculture, as asymptomatic plants can act as hidden reservoirs of infection, leading to unexpected and economically damaging outbreaks in susceptible crops. Addressing this issue, researchers at the National Agriculture and Food Research Organization, Japan, have developed an innovative algorithm to predict the severity of viroid-induced symptoms using unsupervised machine learning[1]. The newly developed algorithm aims to predict the severity of disease symptoms caused by viroids, such as the potato spindle tuber viroid (PSTVd), in host plants like tomatoes. Traditional methods of studying viroid pathogenicity involve biological experiments that are time-consuming and labor-intensive. The new algorithm offers a streamlined approach by relying solely on the genome sequences of the viroids and host plants. The algorithm operates in three main steps. First, it aligns synthetic short sequences of the viroids to the host plant genome. This step mimics the RNA silencing mechanism, a natural defense process in plants that involves small RNAs (sRNAs) targeting and degrading viral RNA. Next, the algorithm calculates the alignment coverage, which measures how extensively the viroid sequences match the host genome. Finally, it clusters the viroids based on this coverage using two advanced techniques: UMAP (Uniform Manifold Approximation and Projection) and DBSCAN (Density-Based Spatial Clustering of Applications with Noise). To validate the algorithm, researchers conducted inoculation experiments, confirming its effectiveness in predicting the severity of viroid-induced symptoms. The results demonstrated a strong correlation between the genome sequences of viroid isolates and the host plants, suggesting that the algorithm can reliably forecast the impact of viroid infections. This predictive capability could significantly enhance control measures against viroid damage and aid in breeding viroid-resistant crops. The study builds on previous research that has explored the complex interactions between viroids and their host plants. For instance, it is known that viroid-derived small RNAs (vd-sRNAs) can induce RNA silencing of endogenous mRNA, leading to severe symptoms in infected plants[2]. However, the exact mechanisms by which viroid infections cause such symptoms remained unclear, partly due to the limited recovery of vd-sRNAs binding specific target mRNAs from infected plants. The new algorithm addresses this gap by providing a predictive model based on genomic data alone, bypassing the need for extensive biological experiments. Previous attempts to develop RNA interference (RNAi)-mediated resistance in plants have had mixed results. For example, transgenic Nicotiana benthamiana lines expressing partial and truncated versions of PSTVd hairpin RNA (hpRNA) were developed to resist PSTVd infection. While these lines showed some resistance, they also exhibited unusual phenotypes resembling viroid infection[3]. This highlights the complexity of viroid-host interactions and the challenges in developing effective control measures. The new algorithm offers a complementary approach by predicting symptom severity without genetic modification, thus avoiding potential unintended effects. Viroids are highly structured, circular, single-stranded RNA molecules that do not code for any peptides but still induce visible symptoms in susceptible plants. The precise mechanisms of viroid pathogenicity remain poorly understood, although it is known that they involve direct interactions with host factors and the production of viroid-specific small RNAs (vsRNAs)[4]. The new algorithm leverages this knowledge by using genomic sequences to predict symptom severity, thus providing a practical tool for managing viroid infections. In conclusion, the development of this machine learning algorithm represents a significant advancement in our ability to predict and manage viroid-induced diseases in plants. By relying solely on genome sequence data, the algorithm offers a rapid and efficient method for forecasting the impact of viroid infections, potentially aiding in the prevention of outbreaks and the breeding of resistant crops. This innovative approach underscores the importance of integrating computational tools with traditional plant pathology techniques to address complex agricultural challenges.

GeneticsBiochemPlant Science

References

Main Study

1) Predicting symptom severity in PSTVd-infected tomato plants using the PSTVd genome sequence.

Published 3rd July, 2024

https://doi.org/10.1111/mpp.13469

Related Studies

2) Potato spindle tuber viroid infection triggers degradation of chloride channel protein CLC-b-like and Ribosomal protein S3a-like mRNAs in tomato plants.

https://doi.org/10.1038/s41598-017-08823-z

3) RNAi mediated inhibition of viroid infection in transgenic plants expressing viroid-specific small RNAs derived from various functional domains.

https://doi.org/10.1038/srep17949

4) Current overview on viroid-host interactions.

https://doi.org/10.1002/wrna.1570


Related Articles

An unhandled error has occurred. Reload 🗙