Georg Kustatscher

Machine learning integration of over 25,000 mass spectrometry experiments yields 15,326 human protein complexes covering 13,769 proteins, plus 511 AlphaFold-derived mutually exclusive subunit pairs.

The paper records proteomes for non-essential gene knock-outs in yeast to uncover principles of protein expression and annotate gene function.

The paper constructs a co-regulation map of the human proteome using mass spectrometry data and machine learning to identify functional relationships between proteins.

The paper shows treeClust, an unsupervised machine-learning algorithm, outperforms correlation-based methods in detecting protein co-regulation in proteomics data due to its robustness to outliers and focus on close linear relationships.

This paper introduces nanoRF, a Random Forest method to detect protein complexes and their relationships in proteomics data, revealing novel functional links and dependencies.