Safety assessment of plant varieties using transcriptomics profiling and a one-class classifier.

Area of Science:

• Agricultural Science
• Computational Biology
• Toxicology

Background:

• Hazard identification for new plant varieties relies on comparative analysis.
• Omics profiling offers an unbiased approach to enhance comparative analysis.
• Classifying novel varieties against a known safe baseline can streamline safety assessments.

Purpose of the Study:

• To explore a one-class classifier for assessing transcriptome profiles of potato varieties.
• To develop a data analysis method for estimating the similarity of novel varieties to a reference class.
• To facilitate hazard identification by prioritizing varieties for further testing.

Main Methods:

• Utilized transcriptome profiles from six potato (Solanum tuberosum) varieties.
• Incorporated data from two growth locations and two harvest years with replication.
• Applied a one-class classifier model to distinguish between 'similar' and 'different' varieties.

Main Results:

• The one-class classifier model demonstrated higher class distances for 'different' varieties compared to 'similar' ones.
• Transcriptome profiling effectively differentiated potato varieties based on genetic and environmental factors.
• The model successfully classified test sets representing 'similar' and 'different' varieties.

Conclusions:

• A one-class classifier approach using transcriptome data can effectively support hazard identification of novel plant varieties.
• This method can reduce the need for extensive biological and toxicological testing by pre-screening varieties.
• The study provides a foundation for a more global and efficient approach to plant variety safety assessment.