Manual for PhenotypeSeeker web tool application
To use the PhenotypeSeeker web tool, first, select the bacterial species and the phenotypes for which the predictions are required.
Next, upload the genomes for prediction.
The genomes could be either in fastA or fastQ format, also they could be gzip compressed, tarballed or both. Tarballing is required to upload more than 20 genomes.
The results will be availabe on the webpage as well as provided in the downloadable text file. If the e-mail address is provided, the results will be also sent by e-mail. User will be also provided by the URL link to return to the results later.
The additional info about the model including performance metrics are provided by clicking the button next to the model. A list of k-mers used in the model as the predictive features is also provided, together with their importance values in the model. This list could be used to discover the genomic variations behind the phenotype.
Example of the results:
| Phenotype | Sample ID | Predicted value |
Estimated probability* |
|---|---|---|---|
| Ciprofloxacin | HUM328 | 0 | 0.44 |
| Ciprofloxacin | ENV126 | 1 | 0.81 |
| Ampicillin | HUM328 | 1 | 0.96 |
| Ampicillin | ENV126 | 0 | 0.08 |
*Uncalibrated
Click here to download the results as a text file
If you require further assistance please contact us via this webpage's contact form or by sending email to erki.aun@ut.ee