User Guide for the READRetro website
This guide provides comprehensive instructions for using all features and options available on the READRetro web interface (https://readretro.net). The official code repository for the web interface is https://github.com/pnucolab/readretro-website.
Viewing and Interpreting Results
Once a prediction job is complete, the “Result Page” will display the findings. This page is designed for clear visualization and interaction.
- Pathway Visualization (Arrows)
Arrows connect compounds in the predicted pathways and are color-coded to indicate the source of the prediction:
Red Arrows: Represent steps generated by the READRetro deep learning models (i.e., novel predictions not found directly in the retrieval databases). This is the default for a predicted transformation.
Blue Arrows: Indicate that the specific single-step retrosynthetic reaction was found in and retrieved from the “Reaction Retriever” database.
Green Arrows: Signify that a sequence of reactions (a partial or full pathway) was retrieved by the “Pathway Retriever,” often from KEGG.
- Information Displayed with Arrows
Additional biochemical context is provided above and below each arrow:
EC Numbers: Enzyme Commission numbers are displayed, classifying the potential enzymes that might catalyze the reaction. These are often clickable, linking to the BRENDA Enzyme Database for detailed enzyme information.
KEGG Pathway Information: If applicable, links or identifiers for relevant KEGG pathways are shown, allowing users to explore the broader biological context. These typically link to the KEGG PATHWAY database.
- Compound Representation (Cards)
Each molecule in the pathway is shown as an interactive card with the following features and color-coding:
Default Card Color: Typically red (or a neutral color) for newly predicted intermediates.
Blue Card Color: Indicates that the molecule is a known compound found in the KEGG COMPOUND database.
Yellow Card Color: Applied when a user clicks on a compound card to highlight it, making it easier to trace its involvement across different suggested pathways.
- Information on Cards:
SMILES Notation: The SMILES string of the compound.
KEGG Entry: If the compound is in KEGG, its KEGG ID is displayed and is clickable, redirecting to the specific entry in the KEGG COMPOUND database.
Structural Diagram: A 2D chemical structure image of the molecule.
- Interactive Features
The Result Page offers several tools to aid in analysis:
Highlighting Compounds: Clicking on a compound card will highlight that compound across all displayed pathways.
Changing Pathway Display Direction: Options may be available to view pathways in the retrosynthetic (product to precursors) or biosynthetic (precursors to product) direction.
- Saving and Re-accessing Results
You can save and revisit your prediction results:
Unique Result Link: Each prediction generates a unique web link. The link follows the format
https://readretro.net/result/followed by your specific [job id] (e.g.,https://readretro.net/result/d0bc0da8ebb245bb87e8aa161fbbd2d6). You can bookmark or share this link to access the specific result page again later.Download Result Button: A “Download result” button allows you to download the predicted pathway information as a .txt file for offline storage, sharing, or further analysis.