{"entity": "researcher", "timestamp": "2026-08-18T14:16:09.888Z", "family": "Nordenstorm", "given": "Olof", "initials": "O", "orcid": "0009-0003-5711-376X", "affiliations": ["Department of Cell and Molecular Biology, SciLifeLab, Karolinska Institutet, Stockholm, Sweden."], "links": {"self": {"href": "https://publications-affiliated.scilifelab.se/researcher/54552421d6f448f9bc2b6bdf93931dfa.json"}, "display": {"href": "https://publications-affiliated.scilifelab.se/researcher/54552421d6f448f9bc2b6bdf93931dfa"}}, "publications": [{"entity": "publication", "iuid": "53deef8658d14591b6b08382650120d7", "links": {"self": {"href": "https://publications-affiliated.scilifelab.se/publication/53deef8658d14591b6b08382650120d7.json"}, "display": {"href": "https://publications-affiliated.scilifelab.se/publication/53deef8658d14591b6b08382650120d7"}}, "title": "Zero-shot prediction of drug responses using biologically informed neural networks trained on phosphoproteomic timeseries.", "authors": [{"family": "Antonopoulos", "given": "Konstantinos", "initials": "K", "orcid": "0000-0003-2781-3872", "researcher": {"href": "https://publications-affiliated.scilifelab.se/researcher/04cb0389161647d888b443050b719c05.json"}}, {"family": "Nordenstorm", "given": "Olof", "initials": "O", "orcid": "0009-0003-5711-376X", "researcher": {"href": "https://publications-affiliated.scilifelab.se/researcher/54552421d6f448f9bc2b6bdf93931dfa.json"}}, {"family": "Nilsson", "given": "Avlant", "initials": "A", "orcid": "0000-0002-9476-4516", "researcher": {"href": "https://publications-affiliated.scilifelab.se/researcher/f2e21dbc1c624f6a841c59e959e948e4.json"}}], "type": "journal article", "published": "2026-03-00", "journal": {"title": "PLoS Comput Biol", "issn": "1553-7358", "volume": "22", "issue": "3", "pages": "e1014100", "issn-l": "1553-734X"}, "abstract": "Cellular signaling is driven by complex, dynamic phosphorylation networks that control growth and survival, and their dysregulation underlies diseases such as cancer. Although modern mass spectrometry enables large-scale quantification of phosphoproteomic responses over time, these measurements remain descriptive and cannot by themselves predict how signaling will evolve under perturbations. Here, we extend a biologically informed recurrent neural network framework (LEMBAS), to learn time-resolved phosphoproteomic trajectories. We introduce two interpretable modules; a phosphosite mapping that links signaling nodes to measured phosphorylation sites and a monotonic time mapping that aligns continuous experimental times to discrete signaling steps. Using synthetic benchmarks and an EGF-stimulation dataset with inhibitor treatments, the model accurately interpolates unseen time points and predicts drug-induced phosphoproteomic responses in a zero-shot setting, outperforming na\u00efve and fully connected baselines. Importantly, the model identifies both canonical and non-canonical signaling effects, including modulation of the transcription factor FOXO3:S7 (from the PI3K/AKT pathway) by drugs affecting PTPN11 (from the RAS/ERK pathway). By combining mechanistic priors with deep learning, our framework provides a scalable approach to interpret and predict dynamic drug responses from phosphoproteomic data.", "doi": "10.1371/journal.pcbi.1014100", "pmid": "41849361", "labels": {"Avlant Nilsson": null, "DDLS Fellow": null}, "xrefs": [{"db": "pmc", "key": "PMC13035234"}, {"db": "pii", "key": "PCOMPBIOL-D-25-02219"}], "notes": [], "created": "2026-08-17T11:12:39.491Z", "modified": "2026-08-17T11:12:39.736Z"}]}