shinyDeepDR

shinyDeepDR predicts anti-cancer drug sensitivity using the DeepDR deep learning model to map tumor or cell-line mutation and gene expression profiles to responses for 265 approved and investigational anti-cancer compounds, supporting precision oncology.

Key Features:

• DeepDR deep learning model: Implements the DeepDR framework for predicting drug sensitivity from genomic data.
• Genomic input types: Uses mutation and/or gene expression profiles from tumors or cell lines as model inputs.
• Find Drug functionality: Predicts sample-specific responses across 265 approved and investigational anti-cancer compounds.
• Find Sample functionality: Identifies cell lines in the Cancer Cell Line Encyclopedia (CCLE) or tumors in The Cancer Genome Atlas (TCGA) with genomic profiles similar to a query sample.
• Per-compound prediction outputs: Produces detailed prediction results for individual compounds to support interpretation of drug–sample interactions.

Scientific Applications:

• In silico anti-cancer drug screening: Enables computational screening of tumor or cell-line genomic profiles against approved and investigational compounds.
• Precision oncology / personalized treatment prediction: Supports prediction of individualized drug sensitivities to inform precision medicine research.
• Translational model selection: Facilitates selection of CCLE cell lines or TCGA tumors with matching genomic profiles for translational studies.
• Investigational drug exploration: Allows exploration of response hypotheses for investigational compounds.

Methodology:

Implements the DeepDR deep learning model using mutation and/or gene expression inputs to predict drug sensitivity across 265 compounds and to query CCLE and TCGA for genomically similar samples.