Here's a suggested Project Description for your PRIDE submission: Project Description This dataset contains mass spectrometry raw files used as training data for SpecFormer, a transformer-based ion intensity prediction model integrated within PatternLab for Proteomics (Spectral Cruncher module). The dataset includes bulk proteomics data from HeLa cells and Mus musculus (C57BL/6) kidney tissue, as well as single-cell proteomics data from WT83 human brain organoids acquired using the cellenOne platform. All samples were analyzed on an Orbitrap Astral mass spectrometer using data-dependent acquisition (DDA). These raw files were used to train instrument-specific models for accurate fragment ion intensity prediction in both bulk and single-cell proteomics workflows. The SpecFormer model and analysis tools described in this dataset are freely available within PatternLab 5.1 at http://patternlabforproteomics.org/51.