Sample Size-Comparable Spectral Library Enhances Data- Independent Acquisition-Based Proteome Coverage of Low-Input Cells

Despite advancements of data-independent acquisition mass spectrometry (DIA-MS) to provide comprehensive and reproducible proteome profiling, its utility in very low-input samples is limited. Due to different proteome complexities and corresponding peptide ion abundances, the conventional LC-MS/MS acquisition and widely used large-scale DIA libraries may not be suitable for the micro-nanogram samples. In this study, we report a sample size-comparable library-based DIA approach to enhance the proteome coverage of low-input nanoscale samples (i.e., nanogram cells, similar to 5-50 cells). By constructing sample size-comparable libraries, 2380 and 3586 protein groups were identified from as low as 0.75 (similar to 5 cells) and 1.5 ng (similar to 10 cells), respectively, highlighting one of the highest proteome coverage with good reproducibility (86%-99% in triplicate results). For the 0.75 ng sample (similar to 5 cells), significantly superior identification (2380 proteins) was achieved by small-size library-based DIA, compared to 1908, 1749, and 107 proteins identified from medium-size and large-size libraries and a lung cancer resource spectral library, respectively. A similar trend was observed using a different instrument and data analysis pipeline, indicating the generalized conclusion of the approach. Furthermore, the small-size library uniquely identified 518 (22%) proteins in the low-abundant region and spans over a 5-order dynamic range. Spectral similarity analysis revealed that the fragmentation ion pattern in the DIA-MS/MS spectra of the dataset and spectral library play crucial roles for mapping low abundant proteins. With these spectral libraries made freely available, the optimized library-based DIA strategy and DIA digital map will advance quantitative proteomics applications for mass-limited samples.

Automated summary

The study presents a sample size-comparable library-based DIA approach, achieving higher protein group identification from small-size library compared to medium-size, large-size, and lung cancer resource spectral library. The approach shows good generality across different instruments and data analysis methods.