A machine learning approach for prediction of mitochondrial proteins in non-model organisms

The evolution of the repertoire of proteins localized to organelles is important for understanding the evolutionary process of organelles. However, experimental methods for identifying organelle-localized proteins have been established only for model organisms and some organisms. Therefore, prediction methods using sequence data obtained from genome and transcriptome analyses, which are relatively easy to obtain, are useful. However, such prediction methods had also been established only for model organisms. In this talk, I will introduce our study in which a machine learning method was used to obtain protein candidates localized to mitochondrion-related organelles in non-model organisms.