I’m Alfred Nilsson, a PhD student working at the intersection of machine learning and proteomics in Lukas Käll’s group within the Department of Gene Technology at KTH. My focus is leveraging transformer models and unsupervised representation learning to decode complex patterns in tandem mass spectra.
My ultimate goal is to show how foundational models, trained on vast datasets, can enhance various downstream tasks. These include improving de novo peptide sequencing from mass spectra, deconvolving chimeric spectra, metric learning for distances between spectra, refining database searching, and potentially sequencing full-length proteins de novo.
If you’re interested in what I’m doing or see a chance for us to work together, feel free to reach out at alfredn@kth.se!