smORFunction

smORFunction predicts functions of small open reading frames (smORFs) and microproteins (<100 codons) to elucidate their roles in biological processes.

Key Features:

• Extensive Database: Aggregates 617,462 unique smORFs compiled from three prior studies.
• Expression Estimation: Re-annotates microarray probes to estimate smORF expression across 173 Gene Expression Omnibus (GEO) datasets.
• Function Prediction Methodology: Uses a speed-optimized correlation algorithm to predict smORF functions by assessing co-expression with genes that have known functional annotations.
• Validation and Success Rate: Validated against literature microproteins and UniProt, with at least one function predicted for 202 of 270 microproteins.
• Comprehensive Coverage: Provides predictions for 526,443 smORFs across up to 265 models spanning 48 tissues or cell types and 82 diseases (including normal conditions).

Scientific Applications:

• Biological Process Elucidation: Enables investigation of smORF and microprotein involvement in processes such as muscle formation, cell proliferation, and immune activation.
• Disease Research: Supports analysis across diverse disease-related datasets for exploring disease mechanisms and potential biomarkers.
• Functional Annotation of Genomic Data: Facilitates functional annotation of genomic and gene expression datasets by linking smORFs to annotated gene functions.

Methodology:

Leverages a database of 617,462 smORFs from three studies, re-annotates microarray probes to estimate expression across 173 GEO datasets, applies a speed-optimized correlation algorithm to infer functions via co-expression with annotated genes, and validates predictions against literature microproteins and UniProt (202/270).