We deliver tailored solutions combining traditional traditional ligand and structure based Computer Aided Drug Design (CADD) methods with AI, Data science and Quantum mechanics to foster your molecule and target discovery programs.
- New library creation. From existing public (private can also be considered) databases as well as from literature mining we are able to molecular libraries that can help you to advance your Drug discovery projects. The library can be target, disease, pathway or population focus as well as a combination, depending on your requirements. The library can be as customized, curated and detailed as you need; for instance including per each compound ADMET properties, synthetic-accessibility, PAINS identification, off-targets information, Nearest Neighbours (similar compounds), development stage, chemical families identification, clusterization or chemical vendors information. Obviously the libraries can be as diverse as you required based on structural or ADMET properties, for instance.
- Curation of a proprietary library. If you already have a proprietary library we can enrich your library with new compounds. We can also also customize and curate it with ADMET properties, PAINs identification, off-targets information, scaffold identification, substructure search or chemical families identification and clusterization. We can also create diverse libraries from it according to your requirements.
- New library creation. Similar to molecular libraries we can also design target libraries for you. The library can be disease, pathway or population focus as well as a combination, depending on your requirements. The library can be as customized, curated and detailed as you need; for instance including existing binding sites together with its conservation in other targets (it could be a source of potential promiscuity and off-targets effects), existing compounds (indicating its development stage), binding patterns and key binding residues, reported mutations, structurally similar targets or its involvement in disease.
- Enrichment of a proprietary database. If you already have a proprietary library we can apply our tools and knowledge to its curation and enrichment; as commented above, binding sites information, existing compounds, binding patterns and key binding residues, reported mutations, structurally similar targets or its involvement in disease can be include in your library.
Ask us what you need and let’s see how can we help you!
We can identify potential targets (and off-targets) for your compounds using different information sources (public/private databases, scientific literatue, etc) and techniques.
- Proficiently mining existing available information in databases and in the literature (or in other sources you indicate or provide us) we are able to get the targets information you need. We customize the searches to fit to your requirements obtaining clean, understandable, human readable data.
If there is not reported information we can predict possible targets using proprietary or third-party ligand based, structure based or hybrid approaches.
Ligand-based approaches :
- Similarity. Using classical or ensemble-based similarity approaches (based on molecular descriptors like structural data, topological information, physicochemical properties or conformational data) we are able to identify the most similar compounds, in the analyzed chemical space, to your molecules. The similarity of a query molecule to a single molecule or a group of molecules allow us to extrapolate that if the reference compound(s) bind to target A, your query molecule can also bind to target A.
- Bioactivity prediction. Using ML models generated to predict certain bioactivity, based on the SAR principle and representative of a single target or protein family, we can elucidate whether or not a query compound could show the evaluated bioactivity. If a compound will behave as a binder of a specific target or protein family or not.
Structure-based approaches :
- Binding (site/mode) conservation. Based on binding sites and binding mode (protein-ligand interactions) conservation we can identified targets to that your compounds can bind. Targets with similar/conserved binding sites or binding interactions. Structural conservation or similarity analysis over binding sites and binding modes databases allow us to identify targets for your compounds
- Virtual Profiling/Inverse screening. The combination of Docking, Molecular Dynamics (MD) and/or Quantum Mechanics (QM) techniques can be used to predict if a molecule can bind to a possible target. Using protein binding sites databases we can evaluate the possible binding of a molecule against hundreds of targets.
Hybrid approaches :
- Network-based. Networks where nodes can be proteins, compounds, or both, with the edges being interactions, similarities, or phenotypic effects can be constructed. Then, using chemical structures and similarities between connections, targets can be identified for query compounds.
- Proteochemometrics. The application of similarity approaches in more complex scenarios where protein descriptors (like sequence or structural/coformational information) are added allow us to identify if a query compound culd bind a certain target or protein family, or even to several targets.
If you already have an identified target or list of potential targets but are unsure of the druggability or if your lead compound would interact with it, we can help you evaluate it.
- Druggability & cavity finding. We can evaluate the potetntial druggability of your target at the same time we analyze it to find plausible pockets, where a drug/ligand can interact. For each pocket different structural features, that alow us to rank them, are calculated.
- Ligand iteraction. Using Docking, Molecular Dynamics (MD) and/or Quantum Mechanics (QM) approaches (depending on the case) we can evaluate if a ligand would be able to bind a target, in general, and a concrete cavity in particular.
Don’t hesitate to contact us if you want to learn more about how you can validate your targets!
Prior to perform most structure-based in silico studies or biomolecular simulations, we have to model the target and/or ligand structures. Probably the most tricky part is to get a corret target structure. Usually 3D structures available in the Protein Data Bank are enough as starting points. If this is the case with few fixes (there are exceptions due to the quality of the resolved structure) your target structure use to be ready to perform a docking, Molecular Dynamics (MD) or Quantum Mechanics (QM) study. We can help you evaluate the quality of the available structures, select the most suitable for your needs and make the necessary fixes to make it ready for further studies.
However it could happen that there are not available structures for your target of interest. If this is your case, don’t worry. We can help you. Using homology modelling techniques (like the nowadays famous AlphaFold ) we are able to predict/generate a reliable 3D structure for your protein, that can be used as starting point for subsequent studies.
Don’t hesitate to contact us if you want to learn how can we help you modelling your target!
Hit identification is a key step in any drug discovery project. Computation can help you to save time and resources finding hit compounds. Nowadays the libraries to screen are in the order of millions, making a proper exploration rather difficult and expensive. Computational approaches, High-Throughput Virtual Screening (HTVS), allows to virtually screen these libraries overcoming these issues. Moreover as a result you will obtain a ranked list of compounds that can be experimentally testing in a reasonable time withotuh wasting resources.
We can help you to find new molecular hits exploring public, private or commercial molecules libraries. The option of use commercial libraries is specially interesting as you can directly order the compounds (we can also take care of this process) and testing them experimentally (we are purely computational but we can also manage to subcontract a third party to do the testing), avoiding any synthesis process. Then we can take the most potent hits and optimize them until become leads matching all your requiements (at this point synthesis used to be required. We can also manage to subcontract a third aprty to do it if needed): new chemical scaffolds, potency improvement, overcaming IP issues, and so on.
To find hits we can use LIgand-based and Structure-based approaches.
Ligand-based approaches. Using known ligands of your protein target, disease, pathway or population of interest, we can find similar compounds to them; compounds with similar molecular properties and/or with a similar behaviour.
- We can screen compounds databases based on molecular descriptors like structural data, topological information, physicochemical properties or conformational data by using similarity approaches.
- We can use previously created QSAR/QSPR models to predict which compounds from a library will behave as known active compounds.
Structure-based approaches. Using the 3D structure of your target of interest by employing in silico binding techniques (mainly docking followed by Molecular Dynamics and/or Quantum mechanics calculations to refine top ranked docking results) we can evaluate if some compounds from a molecular library can bind your target of interest.
- If you know the cavity where the hits are supposed to bind we can evaluate the binding on this cavity. If not we can identified the most plausible cavities on your target, and evaluate the binding in them.
- If key ligand-target interactions are known for your target, we can use them to filter the results, selecting compounds that establish them. Otherwise, we can mine the literature and binding mode databases to identify the key interactions or extract the contacts from the available 3D structures and use them to filter the results
As a general rule, we prefer to combine as many approaches as possible. However, depending on your specific case (available data and so on), we will select the approach that best suits your needs.
Contact us! Together we will devise the best strategy to help you!
Identify a hit is a good starting point for a drug discovery project. However it has to be optimized until become a lead. A compound that can be considered valid to get into (pre)clinical phases. Computationally we can help you to optimize your compounds. Combining human intelligence (our knowledge and chemical/biological intuition) with computational approaches (mainly in silico binding techniques, similarity approaches and QSAR models) we can propose structural modifications (the most favourable according to our computational aproaches) to your compounds to:
- Improve potency. A higher binding strength, as well as optimized binding mode, usually correlate with a higher potency.
- Improve selectivity. We can check the selectivity profle against the target proteins you indicate us.
- Overcome IP issues. We can generate alternative molecules that could be patentable with similar properties to your compounds.
- Improve ADMET and physicochemical properties. We can propose structural modifications that could improve the ADMET and/or physicochemical properties, overcoming potential issues, of your compounds, while maintaining other properties that influence their activity.
If you have a different need, another characterstic to improve, don’t hesitate to conact us. We can talk about it a find a way to solve your problem together.
Contact us to deeply learn how can we help you! Together we will devise the best way to help you!
We can help you to understand how your hits and leads interact with your target protein. We can help you understand which parts of your molecule are key to establish interactions with the target and enhance its activity. We can help you to understand the kind of established interactions and how it correlates with the nature of the aminoacids preent in the bining site. To unerstand wich factors/forces dominate the binding.
This analysis is routinely included in most hit identification and lead optimization services but it can also make as an standalone process. This understanding can also guide us to propose structural modification to improve your compounds ( hit/lead optimization) or to find new compounds that behave as yours (hit identification) or just to deeply understand the Mechanism of Action of your compounds.
Don’t hesitate to contact us if you want to learn more about your ligand-target interactions!
The theory of quantitative structure-activity / property relationships (QSAR / QSPR) is a chemoinformatic approach that allows making predictions of activities and / or properties of chemical compounds. We develop tailored QSAR/QSPR models to to predict different physicochemical, biological and environmental fate properties like:
- Target binding (quantitative, Ki, IC50 or EC50, and qualitative ,activator, inhibitor, agonist, antagonist, PAM or NAM)
- Physicochemical properties (LogP, LogS or pKa)
- ADMET properties (Bioavailability, Blood Brain Barrier permeability, Cytochrom P450 interaction or AMES toxicity)
We don’t believe in QSAR or ML models as black boxes so along with the predictions we provide you the reliability of the prediction based on the applicability domain of the employed model (a prediction can be reliable only if the compound is similar to one or more compounds present in the training set).
Don’t hesitate to contact us about our models! We could find together the best way to help you!
We use Machine Learning, mostly, based QSAR/QSPR models developed by us or from third parties to perform ADMET and physico-chemical predictions of your compounds. We don’t believe in QSAR or ML models as black boxes so along with the predictions we provide you the reliability of the prediction based on the applicability domain of the employed model (a prediction can be reliable only if the compound is similar to one or more compounds present in the training set). The models we use more frequently are listed below (along with other models we may also use). However if you want to predict another property not listed, don’t hesitate to ask us about it. Maybe we already have a model (we regularly try to find new data and construct/get access to new models) or if it is not the case, we can construct a new model to meet your necessities.
- CACO-2 permeability
- Blood Brain Barrier (BBB) permeability
- Plasma Protein Binding (PPB)
- Hepatotoxicity (DILI)
- Intestinal absorption
- Skin permeability
- Cytochrome P450 inhibitor
- Cytochrome P450 CYP2D6/CYP3A4 substrate
- Renal OCT2 substrate
- AMES toxicity
- Acute Toxicity (Rat LD50)
- hERG I/II inhibitors
Do not hesitate to contact us if you want to learn more about the models!
The structure-activity relationship (SAR) is the relationship between the chemical structure of a molecule and its biological activity. Small structural changes can have dramatic effects on the biologic activity. If you have series of compounds or chemical families with measured activities we can help you to understand the structure-activity relationship. To understand detected activity cliffs. To determine which scaffolds are responsible of the activity. To determine which parts of the molecules could be modified to increase the activity or to improve ADMET or physicochemcal properties withuth loosing activity. By using, mainly, in silico binding techniques, QSAR models and/or Matched-molecular pairs analysis we can help you to understand your molecules; help you to disentangle why they behave as they do.
If you have a SAR and you don’t see a clear structure-activity pattern, contact us! Together, we will design the best strategy to help you!
- Drug repurposing, finding new indications for approved or investigational drugs that are outside the scope of the original medical indication, is attracting considerable resources of the current R&D spending and constitutes a determining factor in the life cycle management of pharmaceutical products. Repurposing drug candidates are known for their safety, cost savings, and commercial benefits. That’s because:
- Drugs are considered safe by regulatory agencies. Dosages and secondary effects are already described. Most of the preclincal and Phase I experiments could be avioided.
- Less investment is needed (this will vary depending on the stage and process of development of the repurposing candidate) as in general preclinical, Phase I and Phase II costs will be reduced (while regulatory and Phase III phase costs will be more or less the same as for a usual drug).
Computationally we can help you to find new uses to your compounds or to find a known drug that could act against a ceratin target or disease of your interest using different approaches:
- Molecular similarity. Using similarity approaches (based on molecular descriptors like structural data, topological information, physicochemical properties or conformational data) we can find similar compounds , from a library of experimental or approved drugs, to active compounds against a certain target, protein family, disease or pathway of interest.
- QSAR/QSPR models. Using previously created models we can predict which compounds, from a library of experimental or aproved drugs, could behave as an active compound agaianst a certain target, protein family, disease or pathway of interest.
- In silico binding. Using docking, Molecular dynamics and/or Quantum Mechanics methods, we can screen a library of experimental or approaved drugs over the binding site of the target of interest. Then we can analyze the obtained binding strength and the observed interactions to determine if a compound could be succeptible of acting against the analyzed target.
- Traget/Binding site similarity. If you have a target of interest, we can try to find homologous, structurally similar, targets. Also, after identify and characterize the binding site of your target of interest, we can try to find proteins with homologous biding sites. Known drugs that act against these homologous proteins would be succeptible to act against your target of interest. Using in silico binding techniques we can check this possible association from a comutational point of view.
- Binding/Interactions pattern matching. If you have a target of interest we can characterize the binding/interaction pattern of known active compounds against it. Then we can try to find experimental or approaved drugs that match this interaction pattern. These compounds would be succeptible to act against your target of interest. Using in silico binding techniques we can check this possible association from a comutational point of view.
- Pharmacophore matching. After defining the pharmacophore of compounds acting against your target of interest, we can then try to find experimental or approved drugs that match it. The compounds that do it, would be succeptible to act against your target of interest. Using in silico binding techniques we can check this possible association from a comutational point of view.
- Pathway/Network mapping. Network analysis using genetic, protein or disease data can aid identification of repurposing targets. Then we can interrogate experimental or approved drug libraries to find compounds able to interact with them.
In general we try to combine different approaches ir order to maximize the success rate. However, depending on your necessities as well as on the available data, we will select the approach that better fits you. We customize the searches as much as needed to fit to your requirements.
If you hae a repurposing project in mind, don’t hesitate to contact us! Together we will find the best option to help you!
We can in in silico evaluate the potential effect of a single/several mutations on the stability and the binding capacity of your target. This aprprach can be applied to analyze protein alones as well as protein-protein, protein-peptide and prtein-ligand complexes.
- Computational Alanine scanning. We can mimick experimental alanine scanning experiments by replacing selected amino acids with alanines followed by MD simulations to allow the system to equilibrate and evolve with the new composition. Subsequently, the effect of mutations on the stability of the protein, on the protein and binding pocket conformation as well as on the binding force is studied by means of free energy calculations on protein-ligand / peptide / protein complexes.
- Computational mutations. Instead of Alanines we can also study other Amioacids mutations using the same simulation approach and analysis described above
Contact us if you want to learn about how can we help you to predict mutations effects over your target!
These are the main technniques we believe can help to foster your Drug Discovery project. This is a finite list but we can do more! If you need something not listed here, please do not hesitate to contact us. In the worst case, if we have no expertise on what you need, maybe we have some colleagues that can help you and we will be happy to put you in contact. You have nothing to lose for asking!
We have experience with different targets and diseases. Ranging from Kinases to GPCRs and from Cancer to Neurodegenerative Diseases. Some works related to our experience have been published while other remain confidential by its nature.
- Protein Kinases
- Other Enzymes
- Transcription Factors
Although this is our main expertise, we are open to work with any objective and disease. We are ready for challenges and eager to learn and help.
We combine our own developed workflows and software tools along the years with third party solutions. We use tools recognized by the community and with a proven track record. Depending on the project we select the most suitable options from our portfolio.
However, if you prefer to use an specific tool for you project, we can discuss about it. Even it is out of our usual technology solutions. Complete CADD suites from companies like Schrödinger, OpenEye, BioSolveit or Cresset are not in our usual portfolio but we are open to use them.