SEAS

SEAS performs statistical enrichment analysis of sample sets to annotate metadata neighborhoods and identify shared clinotypes in high-dimensional omics datasets.

Key Features:

• Unsupervised Learning Integration: Leverages unsupervised learning techniques such as clustering and embedding to organize biomedical samples based on multi-dimensional omic profiles.
• Sample-Set Focus: Analyzes groups of biological samples, whether manually curated or automatically clustered, to evaluate shared properties across sets.
• Clinotype Identification: Evaluates sample sets for shared clinotypes including age group, gender, treatment status, and survival days.
• Application to GBM Datasets: Has been applied to combined The Cancer Genome Atlas (TCGA) and patient-derived xenograft (PDX) data to approximate clinical outcomes of radiotherapy-treated PDX samples.
• Clinical Insight Generation: Derives insights from shared clinical measurements within sample sets to inform interpretation of complex biological datasets.

Scientific Applications:

• Omics data interpretation: Interpreting clustered or embedded high-dimensional omics data to uncover clinically relevant patterns and treatment responses.
• Glioblastoma (GBM) translational analysis: Relating sample clusters in TCGA and PDX datasets to clinical outcomes, including radiotherapy-treated PDX samples.

Methodology:

Performs statistical enrichment analysis of sample sets to identify shared clinotypes, using unsupervised learning techniques such as clustering and embedding of biomedical samples.