Session: Parallel session 5 - New topics in OMICs

Fine tuning of ddaPASEF methods improves the immunopeptidomics coverage and throughput, which is further boosted by ML-assisted rescoring

Jeewan Babu RIJAL¹, Aurelie HIRSCHLER¹, Arthur DECLERCQ², Pauline PERDU-ALLOY¹, Charline KELLER¹, Lennart MARTENS², Christine CARAPITO¹

¹LSMBO, IPHC, UMR 7178, University of Strasbourg, Strasbourg, France
²CompOmics group, VIB-UGent - Center for Medical Biotechnology (CMB), Zwijnaarde, Gent, Belgium

Introduction

Identifying ligands for human leukocyte antigen (HLA) molecules, known as immunopeptides, is essential for vaccine and immunotherapy development. LC-MS enables direct identification of those peptides but requires fine tuning from sample preparation to data analysis due to their non-tryptic nature. We refined crucial MS acquisition methods on timsTOF Pro and Ultra. Additionally, we utilized MS²Rescore, an AI tool that enhances identification with confidence.

Methodology

We optimized MS methods on a nanoElute timsTOF Pro and Ultra (Bruker), analyzing immunopeptides from 5.10e8 HL60 cells. Parameters such as ion mobility range (0.6 to 1.75 Vs/cm²), accumulation time (100 to 200 ms), collision energy ramp (10 to 55 eV and 20 to 59 eV), number of PASEF scans (6 or 10), and HLA-specific polygon filters were evaluated. We tested 5 and 25-cm Aurora3 columns with gradient times ranging from 10 to 100 minutes. Peptides were identified using Proline or Sage software and rescored with MS²Rescore.

Results

We benchmarked eight MS methods using ddaPASEF acquisition on timsTOF Pro to develop a protocol tailored for immunopeptides. Optimizing ion mobility, collision energy, accumulation time, and polygon filter increased identified peptides from 4548 to 5793. Sage and MS²Rescore further boosted identification by 77%. HLA-specific polygon filter added another 10%, including 14% singly charged peptides, previously discarded.

The optimized method in Ultra yielded even better performance with reduced injection amount and gradient time. A 100-minute gradient on a 25 cm Aurora3 column identified up to 19376 peptides. Even with a 5 cm column and 10-minute gradient, over 5000 peptides were identified, demonstrating high throughput potential.

Conclusion

Our study highlights substantial advancements in the immunopeptidomics workflow with the timsTOF Ultra, delivering unprecedented performance and throughput. These improvements are promising for identifying neo-antigens in cancer tissues.