Enhancing treatment recommendations in precision oncology with proteomics and phosphoproteomics data

Abstract

While DNA and RNA sequencing are increasingly applied in precision oncology, the therapeutically important (phospho)proteome is rarely included. We implemented a clinical proteomics workflow into two existing molecular tumour board programs. Here, we present the semi-automated analysis of (phospho)proteomic data allowing therapeutic suggestion within a turnaround time of two weeks. Currently, the cohort consists of >500 individual patient samples from various entities, predominantly sarcomas.

2 batches/week (16 patients) are processed with a multiplexing strategy (TMT-11), allowing high throughput and high (phospho)proteome coverage. For most patients, expression data of >8,000 proteins and >20,000 phospho-sites are measured. The quantified proteins and phospho-peptides are normalised within and across batches. A PCA analysis confirmed that most batch effects were successfully removed. The normalised data establishes a background cohort used in calculating z-scores for each protein/p-site.

Single patient proteomes are then analysed at a protein abundance level and phospho activity level using z-scores to find aberrant activity. A strategy integrating different levels of signalling eg. substrate phosphorylation, using manually curated annotations scores 39 cancer-related signalling pathways and 4 treatment-related biomarker groups for relevance. Our results show that (phospho)proteomics can support genomic findings and provide new treatment suggestions.