Bias in resistance gene prediction due to repeat masking. (October 2018)

• Record Type: Journal Article
• Authors: Bayer, Philipp; Edwards, David; Batley, Jacqueline
• Is Part Of: Nature plants. Volume 4:Number 10(2018)
• Page Start: 762
• Page End: 765
• Publication Date: 2018-10
• Subjects: Plants -- Periodicals; 580.5; 580 21
• DOI: 10.1038/s41477-018-0264-0
• Languages: English
• ISSNs: 2055-0278

Abstract

Abstract Several recently published Brassicaceae genome annotations show strong differences in resistance (R)-gene content. We believe that this is caused by different approaches to repeat masking. Here we show that some of the repeats stored in public databases used for repeat masking carry pieces of predicted R-gene-related domains, and demonstrate that at least some of the variance in R-gene content in recent genome annotations is caused by using these repeats for repeat masking. We also show that other classes of genes are less affected by this phenomenon, and estimate a false positive rate of R genes (0 to 4.6%) that are in reality transposons carrying the R-gene domains. These results may partially explain why there has been a decrease in published novel R genes in recent years, which has implications for plant breeding, especially in the face of pathogens changing as a response to climate change. A bioinformatic analysis shows that the variance in resistance gene content in recently published Brassicaceae genome annotations is partially caused by repeat masking, providing implications for plant breeding.