Cells behave like little factories, copying information encoded in DNA to RNA molecules to create a never-ending stream of proteins that come down the production line. However, just as factories occasionally encounter bumps in their conveyer systems, aberrant RNA modifications may also affect protein production with potential impacts on cellular function.
Accordingly, the ability to detect and profile these naturally occurring changes to RNA molecules during gene expression has enormous research and clinical value, given that these changes have been linked to diseases from cancer and heart conditions. Existing approaches to track modified RNA bases, however, are much too complicated to be performed routinely.
“Current methods to profile RNA modifications often involve highly specialized protocols, so only a few teams can perform these experiments,” explained Jonathan Göke, Principal Investigator at A*STAR’s Genome Institute of Singapore (GIS).
To analyze RNA modifications in a simpler and more accessible manner, Göke and his team developed a novel computational technique called xPore, which leverages data from direct RNA-sequencing. By comparing statistically significant differences between multiple samples, xPore can make accurate inferences on where and how many RNA were modified, even without a control sample.
The researchers validated their method in six different cell lines as well as multiple myeloma patient samples in search of one of the most common RNA tweaks known as the N6-methyladenosine (m6A) modification. Indeed, in one cell line, xPore correctly identified over 90 percent of m6A modifications from the top 1,000 positions in the RNA, outperforming current methods in terms of accuracy and efficiency.
Using the xPore technique, scientists estimated m6A modification rates in sequenced wildtype and variant cell types. Next, they validated their results in six different cell lines by testing for the occurrence of the m6A motif.
© A*STAR Research
According to the authors, the technique’s beauty lies in its simplicity. xPore can be fed direct RNA-sequencing data for analysis without the need for matched sample controls, therefore producing results in much fewer steps than current methods. This allows xPore to detect genes that are rarely turned on in the cell by simply pooling samples, creating more opportunities to identify novel RNA modification sites usually missed by approaches requiring a matched control.
Additionally, this computational tool can be extended to extract other genomic insights, which points to xPore’s immense flexibility and potential. “It's even possible to artificially introduce modifications that approximate other biological aspects of interest such as RNA structure. xPore could be used in similar scenarios as well,” said Göke.
Going further, the team is currently developing additional machine learning-based methods capable of identifying RNA modifications using only a single sample. This feature would complement xPore’s current capabilities to unlock novel diagnostic biomarkers of clinical importance.
The A*STAR affiliated researchers contributing to this research are from Genome Institute of Singapore (GIS).
