variant-classification

ACMG Assistant is a student-level, research-oriented variant classification tool developed to explore the practical application of ACMG/AMP 2015 and 2023 guidelines. It combines automated retrieval of annotation data from public APIs with structured interactive evidence collection to support systematic variant interpretation.

BioMed MCP over AlphaFold DB + 8 public sources, with SQLite knowledge graph, offline mode, and explicit clinical-use limits.

Optuna, SHAP ve Streamlit ile genetik varyant sınıflandırma araştırması. Yalnızca kaynak kod ve sentetik demo içerir; TEKNOFEST veri seti paylaşılmaz.

Calibrate functional/in-silico variant scores into ACMG clinical evidence strengths (ClinGen-SVI), with LDLR/AlphaMissense worked example

Machine-learning triage of LDLR variants of uncertain significance using predictor concordance and ACMG-aligned evidence mapping

Privacy-safe, reproducible genomics variant classification pipeline with OOF ensemble validation.

Privacy-safe, reproducible genomics variant classification pipeline with OOF ensemble validation.

ML pipeline for TEKNOFEST 2026 Healthcare AI — preprocessing, feature engineering, model tuning, validation & explainability (code only; competition data excluded).

An open-source Java implementation of the Animal Variant Classification Guidelines (AVCG).

A hybrid quantum machine learning platform for early disease detection

Closed-world grounding layer that reduces LLM hallucination in genetic variant classification. Benchmarked on 163 RASopathy variants.

Points-based ACMG/AMP variant classification (Tavtigian 2020, basis of SVC v4) side by side with the 2015 combining rules — see where they disagree.

Privacy-preserving hybrid RAG pipeline for clinical genomic variant interpretation — local LLM, LangGraph, pgvector, ClinVar, gnomAD, ClinGen, NCCN

Two audits of a RASopathy gene panel: MAVE coverage ascertainment (PS3/BS3) and computational-evidence calibration limits (PP3/BP4)

Variant pathogenicity classifier with calibrated probabilities, gene-grouped cross-validation

TensorFlow neural network models for predicting cancer-associated missense variant pathogenicity using biologically informed feature engineering.

Interpretable ML pipeline (Random Forest / XGBoost / SVM with SHAP) that predicts Poor vs. Intermediate CYP2C9 metabolizer phenotype for diplotypes of indeterminate function, and prioritizes high-impact variants for MD simulation and clinical review.

Pathogenic-vs-benign missense variant classification from anonymized features, under a benign-dominant test prior. Four per-panel models with cross-panel transfer.

ML model to predict pathogenicity and functional effects of genetic variants using ClinVar annotations and pre-computed scores

Machine-learning triage of LDLR variants of uncertain significance using predictor concordance and ACMG-aligned evidence mapping