On the homepage of 3dpredict/Ab you can submit a new job by clicking on the "Start from Sequences Icon"
This will open up a job submission dialog allowing you to provide your own sequences as input and subsequently submit them for calculation. Once you clicked on Start from Sequences you'll see the following dialog:
Job submission dialog
Each section of the dialog will be explained hereafter using the orange indicators of the dialog above.
A: Antibody class
In this section you should select the format of the output antibody structure you wish to calculate properties for and ideally your input sequences should correspond to that format.
- Fv: predict only the variable region of an antibody
- Fab: predict a single Fab region of the input antibody
- Fab2: predict a F(ab')2 fragment instead of a single fab
- VHH: predict variable heavy domain of heavy chain
- scFv: single chain variable fragment is a single sequence construct between the variable light and heavy chain fragments of an antibody
- Ig: predict the full immunoglobulin. If you provide a single Fv sequence as input, 3dpredictAb will complete the sequence to the full Ig automatically and build the second arm as well
- BsAb: bispecific antibody, same as the full Ig but supporting a different Fv on each arm
The choice of the antibody class has an impact on the size of the output structure and as a result also on the runtime of the calculation. The bigger the output structure gets, the longer it runs. In terms of runtime & computational complexity one can sort all classes roughly: scFv, VHH, Fv < VL,VH, Fab < Ig, BsAb
B: Input section
In the input section you can either choose to upload one or multiple files or directly paste sequences.
File Input
In this dialog you can drop one or multiple fasta or Excel files. Fasta input files should be fasta compliant providing valid header and id's ideally. For more information on how to format your fasta files and excel spreadsheets refer to the dedicated page on that topic.
You can use a demo file of 14 sequences available here to get you started.
Direct Input
When activating the direct input form allows you to paste your sequence of choice into the corresponding input fields and additionally add an identifier for the molecule
C: Job Name
Be creative and give an informative job name here, it should be short & clear. This is the identifier allowing to find your results at later stages as well, so it should be easily recognisable by you. For instance your project name / identifier & stage or campaign name.
D: Job Description
Here you can add more details in which context you run the calculation & why with these parameters like at lower pH levels etc.
E: Annotation Scheme
Here you have the choice on how to annotate & number the output model and residue numbers. 5 different choices are available among the most commonly used antibody numbering schemes.
The Annotation scheme has an impact on two things in the output of the predictions, the aminoacid sequence numbering in the tables & the structure and on the defintions of CDR ranges on the sequence.
NB: The annotation scheme IMGT-sequential uses a primary amino acid sequence numbering and IMGT for CDR range definitions. CCG-sequential uses a primary amino acid sequence numbering and CCG for CDR range definitions.
F: pH
By default you'll run the property calculation at pH 7.4, but you can change to any valid pH here. Note, that 3dpredictAb will run an ensemble of property calculations on conformations around the pH of your choice. For pH 7.4 for instance a pH range from 6.4 to 8.4 is sampled with 100 conformations in total to extract the final descriptors and 95% confidence intervals.
G: Salt type
If you want to change the salt used for the simulation from NaCl to KCl you can do that here
H: Salt concentration
Here you can adapt the salt concentration used for the simulation to match your real life conditions
I: Runtime & Cost
Once you uploaded your sequences and set the desired antibody format you'll see a cost & runtime projection here. This allows you to have an approximative estimate of how long your calculation will run and how many credits will be consumed.
For more information on credits you can refer to the credits section of this documentation