09/16/2026
Single-cell proteomics coverage ceiling broken. Most published single-cell studies identify a few hundred to ~2,000 proteins per cell β limited by how DIA fragments and how missing spectra are handled.
A new Genome Biology paper introduces Full-DIA, a deep learning approach that substantially improves proteome coverage, quantitative accuracy, and analysis speed for single-cell proteomics on diaPASEF platforms. The key: train fragment prediction on diaPASEF data structure itself, so missing spectra are recovered rather than discarded.
π What changes for researchers:
β Single-cell heterogeneity studies get far more proteins per cell β closer to bulk coverage at single-cell resolution. β Cell-type identification in complex tissues becomes more reliable with broader protein signatures. β Paired multi-omic studies (where the proteome was the bottleneck) become practical.
Pipeline published openly β any lab with timsTOF access can adopt it.
π Read the full paper: https://doi.org/10.1186/s13059-026-04087-x
diaPASEF improves ion utilization and sensitivity by synchronizing quadrupole isolation with trapped ion mobility separation, making it suitable for single-cell proteomics. We present Full-DIA, a deep learningβdriven software that enhances proteome coverage, quantitative accuracy, and analysis spe...