RNA editing is a natural process that alters RNA molecules after they’ve been created from DNA. One common form of RNA editing is called adenosine-to-inosine (A-to-I) editing, where certain adenosine (A) bases are changed to inosine (I), which cells treat as guanine (G). This process can change how genes are expressed, potentially affecting health and disease.
However, detecting A-to-I RNA editing can be tricky. DNA mutations (permanent changes in the DNA sequence) and technical errors during RNA sequencing can make it hard to tell what’s truly an RNA edit versus a mutation or mistake. To address this challenge, researchers at the Children’s Hospital of Fudan University have developed a new computational tool called DEMINING.
Developing DEMINING embedded DeepDDR model forDNA mutations (DMs) and RNA editing sites (REs) classification
a Construction of a stepwise DEMINING computational framework for direct DNA mutation (DM) and RNA editing (RE) classification. HPB hits per billion mapped bases, MF mutation frequency, MR mutation read. b Schematic diagram of an embedded DeepDDR model for DM and RE classification. Left, features extract strategy by the co-occurrence frequencies of each mutation site with its context bases (CMC). Right, DeepDDR model architecture. c Evaluation of different models on RE identification. Receiver operating characteristic (ROC, left) curves and precision recall curves (PRC, right) of DeepDDR (red), EditPredict (purple), and RED-ML (blue) were shown to indicate their performance on RE identification with the test set. Area under ROC (AUROC) and area under PRC (AUPRC) values of DeepDDR (red), EditPredict (purple), and RED-ML (blue) were included in the figure. d Evaluation of DeepDDR on DM identification. ROC (left) and PRC (right) of DeepDDR were shown to indicate its performance on DM identification with the test set. AUROC and AUPRC values of DeepDDR were included in the figure
What is DEMINING?
DEMINING is a step-by-step computational framework designed to help scientists more accurately identify and differentiate between RNA editing sites and DNA mutations using RNA sequencing data. It includes a deep learning model called DeepDDR, which helps the system “learn” to distinguish between the two types of changes.
This system is important because RNA sequencing data often contains a mixture of signals from both RNA edits and DNA mutations. Without the ability to separate them correctly, it’s hard to make accurate conclusions about gene function or disease mechanisms.
How DEMINING Works
DEMINING uses a multi-step process that carefully analyzes RNA sequencing data to identify potential RNA editing sites and DNA mutations. The deep learning component, DeepDDR, helps refine this identification process by improving its accuracy. After undergoing transfer learning (a method that helps the model adapt to different types of data), DEMINING can also be applied to non-primate samples, making it versatile for various research studies.
Application in Cancer Research
DEMINING was applied to samples from patients with acute myeloid leukemia (AML), a type of blood cancer. The tool was able to identify new DNA mutations and RNA editing sites that had previously been overlooked. Some of these changes were found to be associated with increased expression of certain genes, which could be important in cancer development or progression. Additionally, DEMINING discovered RNA editing sites linked to the production of neoantigens—molecules that can trigger the immune system to attack cancer cells.
Why This is Important
By accurately distinguishing between RNA editing and DNA mutations, DEMINING opens up new possibilities for understanding complex diseases like cancer. The discovery of new RNA edits and mutations in AML could lead to the identification of new biomarkers or therapeutic targets, potentially improving how we diagnose and treat cancer.
DEMINING represents a powerful tool for researchers to better understand the intricate processes of RNA editing and DNA mutations, especially in the context of diseases like cancer. This advancement could lead to more accurate data analysis, furthering our knowledge of gene expression and disease mechanisms, and paving the way for future breakthroughs in personalized medicine.