Input must be a secretome. EffectorP 3.0 is trained to separate effectors from other secreted proteins.
Never submit a whole proteome — the cytoplasmic signal is also recognised in non-secreted intracellular proteins,
which would flood the result with false positives. Run SignalP/Phobius (signal peptide) + TMHMM (transmembrane domains)
first, and keep proteins that have a signal peptide and no transmembrane domain.
Each protein is scored by 60 WEKA models, so runtime grows with the number of sequences. Please do not close this page.
Prediction results
Command line and logs
⚖
Submit a FASTA file of secreted fungal/oomycete proteins to predict which are apoplastic or cytoplasmic effectors.
Recommended workflow
Predict the secretome. Take the pathogen proteome and keep proteins with a signal peptide and no transmembrane domain
— e.g. SignalP 6.0 -t euk or Phobius for the signal peptide, plus TMHMM 2.0 to exclude TM domains.
The original EffectorP 3.0 protocol used SignalP 4.1 -t euk -u 0.34 -U 0.34 followed by TMHMM.
Submit the secretome here. EffectorP 3.0 classifies each protein as an apoplastic effector, a cytoplasmic effector,
a dual-localised effector, or a non-effector, with an ensemble probability.
Prioritise candidates. The authors recommend overlaying infection-stage RNA-seq expression or structural similarity to
known effectors on top of the EffectorP probabilities — the probabilities themselves are rough estimates and should not be over-interpreted.
Citation & implementation
Sperschneider J, Dodds PN. EffectorP 3.0: Prediction of apoplastic and cytoplasmic effectors in fungi and oomycetes.Molecular Plant-Microbe Interactions 2022;35(2):94–99.
doi:10.1094/MPMI-08-21-0201-R
· source code (GPLv3)
Server-side implementation: EffectorP 3.0 with WEKA 3.8.4 on OpenJDK 17. EffectorP is not a secretome predictor
and is not applicable to bacterial, nematode or insect effectors.