Developing machine learning methods for graph-structured data is a theme of paramount importance in bioinformatics. Indeed, graphs are the structure of choice for representing complex objects such as molecules and proteins, which have a central role in many biological problems, as well as for capturing the interactions among them, such as in protein-protein interaction networks. This PhD project is aimed at developing graph learning methods for biomarker discovery from proteomics data, including protein sequences, structures, protein networks and protein abundances. The use of machine learning methods in this field could have a beneficial impact on crucial topics in biology such as disease diagnosis and drug discovery. Towards this end, we will explore, develop and refine learning methods based on graph kernels, significant pattern mining, graph neural networks and topological data analysis.