The Enlace Bio virtual screening service
How our system works
Fasta seq: MKAPLLNGECKAVLFTLVQRSAPGIL…
Step 1: Receive molecule and protein sequence
Step 2: Cofold molecule with protein
Step 3: Assess affinity strength on 3D structure
IC50: 100-1000 nM
(estimated range)
Step 4: Return affinity strength range
How we work with our clients
1 Target selection
The client specifies the protein target for which they are interested in doing a virtual screening.
2 Virtual assay validation
We assess the likelihood of success in a pre-study. We review these results together and proceed only if we have high confidence in the potential success of the project.
3 Virtual screening
We then perform the virtual screening, either on a large molecule library or on a list of molecules provided by the client. We first leverage an AI system to model the 3D structure of the protein and ligands as well as analyse the affinity strength. Hits are confirmed using traditional methods to maximise likelihood of binding in the subsequent in vitro assay.
4 Results delivery
We deliver the screening results to the client, providing them with a shortlist containing the most promising molecules for further investigation and lead optimisation. This diverse selection of starting molecules maximises their chances of success in the subsequent lead optimisation stage.
Our technological edge
Structure-based analysis
Accurate binding assessment requires the 3D structure of the protein–ligand complex. Our system uses advanced AI methods to simulate the docked structure, whether starting from the amino acid sequence or the protein structure. If a bound 3D structure is found, the system analyses it to determine the molecule’s binding strength.
Works for novel targets
By leveraging the latest AI technologies the system works even when there is no experimental structure data available for the protein. This unlocks virtual screening for a new set of targets previously unreachable with traditional docking based virtual screening methods.
Validating hits with traditional methods
While AI technologies are still maturing and occasionally hallucinate hits we confirm all hits we find with traditional methods before returning any results to the client. This further increases the fraction of virtual screening hits that can be confirmed in the lab.
More data
Our AI system is trained on hundreds of thousands of data points containing experimentally determined affinity strength and computer-generated 3D structures. This is over 10x more data than used by academic virtual screening methods.