How to Launch an Alpha Fold model

Step 1. Open the Alpha Fold job. To begin, navigate to the LENSai portal. You can do this by either accessing the homescreen, where you'll find various options available, or by going through the settings menu. If you prefer a more direct approach, you can select the "jobs" section and then proceed to "job overview." This will provide you with a comprehensive list of your ongoing and completed jobs. Once you are in the appropriate section, locate the Alpha Fold job you wish to open.

Step 2. Click "RUN JOB" on the AlphaFold+ icon

Step 3. Click "Select dataset or file" to select a pre-uploaded dataset from the 'input_dataset_id' field. When you click on the input field, a dropdown menu will appear, showcasing all available datasets that have been previously uploaded to the system. Take a moment to review the list and identify the dataset that best suits your analysis needs. It's important to note that you should select the entire dataset rather than individual files contained within it, as the process is designed to work with datasets as a whole. To aid in your selection, you may see additional details such as dataset names, descriptions, or timestamps indicating when they were uploaded. Once you have made your selection, confirm your choice to proceed to the next steps in the Alpha Fold analysis workflow.

Step 4 Click the "Search" field to find a certain data file 

Step 5. Tweak the Alpha Fold Settings:

Here's a detailed breakdown of the AlphaFold settings:

1. Model Preset (--model_preset):

   • Monomer: Suitable for predicting the structure of single protein chains.

1. • Monomer PTM: Includes predicted alignment error and pLDDT confidence scores for assessing the accuracy of specific regions.
   • Multimer: Designed for protein complexes involving multiple chains or interactions.

2. (Only for AlphaFold 2.3.0) Database Preset (--db_preset): 

   • Full Databases: Uses large databases for improved prediction accuracy, requiring more computational resources.
   • Reduced Databases: Uses smaller databases, trading accuracy for faster execution and less memory usage.

3. Models to relax (--Models_to_relax):

   • Enables you to choose which models should undergo relaxation steps, with options for All, Best, or None.

4. Limit the range of MSA (--msa_range):

   • Adjust the preset MSA range for generating the quality report. This setting allows users to focus on specific sequence portions of interest, with a narrower range enhancing analysis accuracy by targeting relevant data. Conversely, a broader range offers a comprehensive overview, beneficial for understanding structural context. Balance the selected range with available computational resources, as larger ranges may demand more processing time and memory.

Step 6 Provide an experiment name. This name serves as a unique identifier for your immunogenicity analysis, making it easier to reference and track your work later. When creating your experiment name, it is crucial to avoid using white spaces, as the current system does not support names with spaces. Instead, consider using underscores (_) to separate words for clarity. For example, instead of naming your experiment "Structural Study 2023," you could use "Structural_Study_2023". 

Step 7. Launch the calculation by clicking the run job button at the bottom left. After confirming your selections and entering all required information, you are ready to initiate the analysis. Clicking the run job button will start the process, and you will receive a notification or progress indicator informing you of the job's status. Depending on the complexity of your analysis and the size of your dataset, this process may take some time. You can monitor the progress through the interface and will be notified once the analysis is complete, enabling you to review and interpret your results efficiently.