EffectorO

If you have any questions or trouble running this, please email authors at [email protected] or [email protected]

Predicting oomycete effector genes using lineage-specificity and machine learning classifiers

Please see our EffectorO bioRxiv preprint for more details! And see our EffectorO-ML web server that is available.

Data used

• For secretomes, EffectorO-ML hits, EffectorO-LSP hits, WY domain-containing proteins and RXLR-EER domain-containing proteins from the 28 genomes Oomycete genomes, see the EffectorO_genome_Results directory.
• For the original ORF files used, see this google drive link: https://drive.google.com/drive/folders/1iFH0nOd__SOluOQa4eRSVNCE0kEhQLu-?usp=sharing

How to get machine-learning predicted effectors from an oomycete fasta file (works best on secreted proteins):

Using the EffectorO-ML web server:

1. Navigate to https://effectoro.onrender.com
2. Upload the FASTA file of our choice, of predicted amino acid sequences
3. See your results in the data table! Can sort each column by clicking the arrows, and can search for sequences by the ID

Using the EffectorO-ML command-line tool:

1. make sure python3 is downloaded

python3 --version

2. git clone this repository, then make sure all packages are downloaded

python3 -m pip install --user -r requirements.txt

3. cd into scripts directory to run ML prediction pipeline

 cd machine_learning_classification/scripts

4. run this on command line:

python3 predict_effectors.py YOUR_INPUT_FASTA_PATH

• output:
  • csv of IDs|class_prediction|meaning|probability_of_prediction
  • fasta file of predicted effectors