1. Fold sequences

Protein folding is the process in which an amino acid chain adopts its functional three-dimensional structure. In Fastfold, you can carry out this process accurately and efficiently by combining automation and interaction with advanced models.

Below, we will cover what you need to know before moving on to a step-by-step practical example.

Where is the "Fold" feature located?

To find the feature, go to the sidebar, navigate to the "Workflows" section, and click on "Fold."


What types of sequences does Fastfold accept?

Fastfold currently supports four types of sequences:

  • Protein

  • DNA

  • RNA

  • Ligand

The user can select the type of sequence they want to fold.


How do I enter my sequence?

There are multiple ways to enter sequences in the "Fold" feature. You can manually add fields and fill them in, allowing you to input one or multiple sequences at a time. You can also upload .fasta or .fa files, which allow you to submit multiple sequences at once.

1. Manually Uploading Sequences

You can add one or more fields to enter sequences by clicking the "Add" button to include additional fields.


2. Uploading .fasta or .fa Files

Follow these steps to upload one of these files:

1

Click the "Upload" button located next to "Add."

2

Your browser's file explorer will open, and you will need to select a .fasta or .fa file. Make sure to keep your files in easily accessible locations for quick access.

3

Once you have selected the file, click "Open" in the file explorer.

4

When you upload the file, multiple fields will be created automatically, each containing a sequence from the file.

Remember that the sequences uploaded at this stage can be edited. You can remove any sequences you don’t want to include in the request or manually edit them as needed.

If you want to test this feature, you can find a sample file here:


Model Type Selection

For sequence execution, you can select one model to run the process. Model availability depends on the type of sequence you want to fold. Below are the available models for each sequence type:

Secuence type
Models

Protein

Boltz - 1 Monomer-ptm-colabfold Multimeter-v3-colafold solosep-esm-meta

DNA

Boltz - 1

RNA

Boltz - 1

Ligand

Boltz - 1


Confirmation Modal

When executing one or more sequences, a confirmation modal will appear with additional information and customizable options based on your needs. In this modal, you can configure:

  • Job name

  • Total sequences and Mean length overview

  • Toggle option: "Relax prediction"

  • Toggle option: "Use custom seed"

  • Details of the sequence(s) to be executed

The objective of each of these is:

1. Job name

You can name the job that will contain the sequence execution.

2. Info: Total Sequences and Mean lenght

The Total sequences indicates the total number of protein sequences included in the current job, while the Mean length provides the average length of these sequences. This overview helps you assess the size and complexity of the sequences being processed.

3. Relax prediction

Relax Prediction is an option that adjusts the protein's structure after folding, reducing strain and improving stability. Enabling it helps obtain a more accurate and functional model.

4. Use Custom Seed

Use Custom Seed is an option that allows you to set a specific value as the starting point for the prediction. This helps ensure reproducible results, as using the same seed enables Fastfold to generate consistent structures across different runs.

5. Details

En esta sección, encontrarás un resumen con la información clave de la predicción. Aquí podrás ver la proteína que estás plegando junto con la configuración aplicada, incluyendo el valor de Mean Length, que representa la longitud promedio de la estructura generada. Esto te permitirá revisar y confirmar los parámetros antes de continuar con el análisis.


Now, let's go through a step-by-step practical example of protein folding.

Last updated

Was this helpful?