In a study published in the journal Nature Methods, researchers from the Agency for Science, Technology, and Research and the National University of Singapore have described software that accurately predicts chemical alterations of RNA molecules based on genomic data. The researchers pointed out that the software they created was able to detect more than 160 RNA modifications.

The most common modification identified by the model was N6-methyladenosine (m6A), which has been associated with several human diseases, especially cancer. Previous techniques were unable to find m6A at the single-molecule level, which is essential for comprehending the molecular pathways involving m6A. Mr. Christopher Hendra, the study′s first author and a current Ph.D. student at A*STAR′s Genome Institute of Singapore (GIS) and the NUS Institute of Data Science, explained that in traditional machine learning, each set of data must be labeled differently, and the problem with recognizing m6A is that there is a large amount of data with ambiguous classifications.

However, the team circumvented these constraints by utilizing direct nanopore RNA sequencing, a cutting-edge method that sequences an unmodified raw RNA molecule along with its RNA modifications.

In this study, they created m6Anet, a software program that uses multiple-instance learning (MIL) and extensive nanopore RNA sequencing data to train deep neural networks to precisely detect the presence of m6A. They added that m6Anet is capable of precisely predicting the presence of m6A from a single sample across species at the single-molecule level.