Gene Inspector Pro now includes two new features for genetic analysis: AlphaMissense and DANN scores. They are designed to give you deeper insights into your genetic data and help you understand the potential impact of genetic variants.
AlphaMissense is an AI model from DeepMind that predicts how likely a missense variant is to be pathogenic. Unlike traditional methods that rely on existing clinical data, AlphaMissense leverages protein structure and evolutionary conservation to make its predictions. This allows it to provide unbiased insights, especially for rare or novel variants. Read the original published paper on AlphaMissense.
AlphaMissense analyzes how a single amino acid change might affect a protein's 3D structure and compares it to similar changes across different species. This approach, which doesn't rely on human-curated databases, reduces bias and improves prediction accuracy.
You can find AlphaMissense scores under the "AMS" label in the variant details. There is also a new panel, "High AlphaMissense Score," which gathers all variants with a score of 0.5 or greater.
DANN (Deleterious Annotation of genetic variants using Neural Networks) is another powerful tool for predicting the disease-causing potential of genetic variants. It uses a deep learning model to analyze a wide range of genomic features and assign a score from 0 to 1.
A DANN score of 1.0 indicates a high likelihood of being harmful, while a score of 0.0 suggests the variant is likely benign. A score greater than 0.96 is generally considered to be a strong indicator of a potentially harmful variant.
A new panel for DANN scores is also available, helping you to quickly find variants with high DANN scores for further investigation.
These new features are a valuable addition to the toolset for genetic analysis.