New version of the SignalP signal peptide predictor
A new version of the widely used SignalP predictor has been published in Nature Methods, as a result of a collaboration between Danish researchers and Gunnar von Heijne at DBB. The new method is based on Protein Language Models, a powerful machine-learning approach that has not previously been applied in this field. SignalP 6.0 has been tailored to predict five different kinds of signal peptides and can be applied to metagenomic data of unknown species provenance.
Last updated:
January 4, 2022
Page editor:
Alexander Tuuling
Source: DBB
