Troubleshooting

Why is my protein binder generation taking longer than expected?

Extended times occur for large/complex targets, high number of binders, or during peak usage

Can I close the platform and return to my projects later?

Yes. Feel free to close your browser and return to your project(s) later. Generation and scoring will continue.

Can I run multiple sessions at the same time?

Yes. You can generate and score binders in multiple sessions/for different targets at the same time.

My generation failed. What should I do?

• Verify target protein structure is complete and valid
• Check binding site accessibility (refer to the tips in Getting Started)

The platform seems unresponsive. What should I do?

• Try refreshing your browser. There may be some cases during the early access phase, where user logins time out after 1 hour. Your projects will be unaffected if this happens. Simply log back in and continue designing.

The generated binders don't look reasonable. What went wrong?

• Common causes: the selected binding hotspots are not optimal for that target or the chosen binder length may not be suitable for the type of binder (for example macrocycles are typically no longer than around 40 residues)
• Our model is still improving. If you have feedback about its performance that would be helpful, please share it at feedback.latentlabs.com

Why can I not generate binders that pass the in-silico filter?

• There could be a number of reasons for why you may not be able to generate passing binder designs in your current settings.
• The expected computational success rate of Latent-X is dependent on the target and the hotspots that are used and typically varies between 1% and 30%. Therefore it is advisable to change the hotspots and / or increase the number of designs that you are generating. Refer to the tips in Getting Started.
• It is also worth highlighting that designs that are close to the passing thresholds of the computational filter may also be valid binders in the lab.
• If you continue to be unable to make designs, we would love to learn of the case and please fill in our feedback form atfeedback.latentlabs.com.