FABIAN-variant

FABIAN-variant predicts the impact of non-coding DNA variants on transcription factor binding to assess potential regulatory effects on gene expression.

Key Features:

• Transcription Factor Flexible Models (TFFMs) and Position Weight Matrices (PWMs): Integrates 1224 TFFMs and 3790 PWMs to predict transcription factor binding changes.
• HMM-based TFFMs: TFFMs are based on hidden Markov models and capture positional dependencies for improved accuracy over traditional PWMs.
• Transcription Factor Coverage: Evaluates the effects of variants on 1387 human transcription factors.
• Model Integration: Combines results from multiple models to produce nuanced predictions of binding-affinity changes for each variant–transcription factor pair.
• Promoter Region and Regional Filtering: Focuses analyses on promoter-region variants and can restrict searches to variants near candidate genes.
• Input Formats: Accepts variant inputs including VCF files generated from high-throughput sequencing.
• Implementation: Implemented in C++ for computational speed and efficiency.

Scientific Applications:

• Regulatory impact analysis: Predicts changes in transcription factor binding to infer potential gene expression alterations.
• Disease and phenotype investigation: Aids identification of genetic contributors to diseases and phenotypic variation.
• Functional genomics: Supports detailed analyses of promoter-region variants to study regulatory mechanisms.
• Personalized medicine: Informs variant interpretation relevant to individualized genetic assessments.
• Evolutionary biology: Assists studies of regulatory variation and its evolutionary consequences.

Methodology:

Integrates 1224 TFFMs (HMM-based) and 3790 PWMs and combines model outputs to predict binding-affinity changes for variant–transcription factor pairs; implemented in C++.