In iTHEMS biology seminar on February 12, Keitaro Kume (Univ. of Tsukuba) gave us a talk about applying machine learning to the analysis of non-model organisms that are difficult to experiment with. First, Dr Kume introduced us about the brief mitochondrial evolution and several eukaryotes possessed not typical mitochondria as like Mitochondrion related organelle (MRO) which is highly degraded mitochondria. Next, he described the detection of mitochondrial localization signals and their applications, explaining that it is difficult to detect localization signals in mitochondria of non-model organisms, especially MROs. Finally, He obtained a large amount of data on MRO-localized proteins in non-model organisms, and through machine learning using the data as training data, he created a detector that can detect MRO-localized signals in such non-model organisms. He presented a good example of the connection between biological evolutionary research and machine learning, which led to a lively discussion at the seminar. Thank you very much, Kei!