Back

predatoR: an R package for network-based mutation impact prediction

Gurdamar, B.; Sezerman, O. U.

2022-11-30 bioinformatics
10.1101/2022.11.29.518310 bioRxiv
Show abstract

MotivationClassification of a mutation is important for variant prioritization and diagnostics. However, it is still a challenging task that many mutations are classified as variant of unknown significance. Therefore, in silico tools are required for classifying variants with unknown significance. Over the past decades, several computational methods have been developed but they usually have limited accuracy and high false-positive rates. To address these needs, we developed a new machine learning-based method for calculating the impact of a mutation by converting protein structures to networks and using network properties of the mutated site. ResultsHere, we propose a novel machine learning-based method, predatoR, for mutation impact prediction. The model was trained using both VariBench and ClinVar datasets and benchmarked against currently available methods using the Missense3D datasets. predatoR outperformed 32 different mutation impact prediction methods with an AUROC value of 0.941. AvailabilitypredatoR tool is available as an open-source R package at GitHub (https://github.com/berkgurdamar/predatoR). Contactberkgurdamar@gmail.com

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.