Comprehensive annotation and analysis of human microproteins by human microprotein atlas platform

Small open reading frames (sORFs) and their encoded microproteins are increasingly recognized as a widespread but poorly characterized component of the human proteome. Here we show a Human Microprotein Atlas (HMPA) platform (http://www.cuilab.cn/microaf) that integrates 617,462 human microproteins with sequence features, predicted three-dimensional structures, variant effect predictions, subcellular localization, essentiality, and bioactivity annotations. Structural analyses revealed substantial conformational diversity, and structure-based functional inference suggested roles in cellular regulation, signaling, and cell-cell communication. Deep learning-based analyses further prioritized microproteins with tissue-specific essentiality and enabled the prediction of diverse bioactivities across multiple functional categories. To support exploration of these data, we also provide an interactive web server with sequence-, structure-, and function-oriented search and analysis utilities. This resource provides a systematic framework for studying microprotein biology and supports the discovery of functional and therapeutically relevant microproteins.