PinPor

PinPor predicts the pathogenicity of small insertions and deletions (INDELs) of up to 21 base pairs in human genetic sequences by analyzing their effects on RNA-binding protein (RBP) affinity and sequence features to identify variants likely to cause inherited diseases; small INDELs account for 18% of recorded mutations causing inherited diseases and are present in 24% of documented Mendelian disorders.

Key Features:

• Data Integration: PinPor leverages disease-causing INDELs from the Human Gene Mutation Database (HGMD) and neutral INDELs from the 1000 Genomes Project for model training.
• RNA-Binding Protein (RBP) Affinity Analysis: The method evaluates how INDELs alter RBP binding, addressing effects relevant to post-transcriptional regulation and alternative splicing.
• Sequence Feature Identification: PinPor assesses sequence features that distinguish pathogenic from neutral INDELs, including proximity to splice sites and potential impacts on RNA secondary structure.
• Machine Learning Model: The tool employs a machine-learning algorithm trained on RBP-affinity changes and sequence features to predict INDEL pathogenicity.

Scientific Applications:

• Pathogenicity prediction of novel INDELs: PinPor prioritizes newly observed small INDELs for further study by predicting their likelihood of being disease-causing.
• Mechanistic interpretation of Mendelian disorders: The analyses elucidate how INDELs may disrupt post-transcriptional regulation, splicing, and RNA structure to inform molecular interpretation of inherited disease.
• Predictive performance for variant interpretation: The model reports a Matthews correlation coefficient (MCC) of 0.51 and an accuracy of 75%, supporting its use in genetic research and variant interpretation.

Methodology:

Collates INDELs from HGMD and the 1000 Genomes Project, analyzes effects on RBP-binding sites and sequence features (including splice-site proximity and RNA secondary structure), and trains a machine-learning model on these parameters to predict pathogenicity.