English
Español
Frequently asked questions
In this section we add progressively the questions that our collaborators and clients transmit to us in relation to our methods and predictive models.
ProtoQSAR is a consultancy company or a vendor of specialized software on molecular modeling and chemoinformatics?
ProtoQSAR is a company that conducts studies for its clients and collaborators, currently we do not sell molecular simulation programs, nor our QSAR models. We provide our services to groups and entities that for the most part do not have specific training in chemistry-computing and molecular modeling -or it is very limited- and therefore our expertise can represent a greater added value for them.
How long does it take for a client to get the results of his project?
The projects in which we work are very different from each other, and prior to their execution we make an evaluation of the time necessary. The great advantage of computer studies compared to experimental tests is that we can obtain results in a much shorter time. For example, the prediction of the profile of a compound by QSAR methods is very fast (few days), since we already have models for this kind of evaluation. The virtual screening of a database to search for a candidate interacting with a particular target can represent 1 week to 2 months, depending on the number of chemicals filtered (a few thousands or hundreds of thousands).
Can I propose a collaborative project to ProtoQSAR? How?
At ProtoQSAR we are very interested in the development of new collaborative projects in which chemoinformatics and molecular modeling can provide an added value to our collaborators. If you have any idea to raise, please contact us explaining your proposal, and in a very short time we will give you an answer.
How much do ProtoQSAR services cost? How are the costs estimated?
Costs vary depending on the type of service we provide and the time we spend on a project. Since computing allows results to be obtained in a much shorter time than experimental tests, our services are also much cheaper.
Computational design
What software does ProtoQSAR use to develop its models?
In ProtoQSAR we work with the whole palette of molecular modeling and chemoinformatics techniques, and therefore we have all the necessary programs for the execution of our projects. We systematically use standardized programs that are well referenced in the scientific literature, and sometimes we carry out our own software developments for specific needs. In any case, these developments remain the property of ProtoQSAR, and are not for sale.
What degree of confidence have the computer models?
Computational studies are not an “exact science” and a certain degree of uncertainty is implicit, as well as it occurs with experimental tests (both in vitro and in vivo) to a greater or lesser extent. We conduct our studies according to scientifically recognized quality guidelines, and our extensive experience (demonstrable through our scientific and technical publications) makes our studies as predictive as possible. Before conducting a study, we inform our clients about the reasonable results’ expectations. At the time of providing the results, we also contribute our expert evaluation and inform to the customer of the potential limitations of the obtained models.
How much time is needed to do a virtual screening? And how many compounds can be filtered?
Both questions are linked: the execution time of a virtual screening depends essentially on the number of compounds that are screened. To give an order of idea, an already developed QSAR model allows to evaluate tens of thousands of compounds in a single day. The screening by docking (in a therapeutic target with known 3D structure) require a longer time, but we can also evaluate thousands of compounds in a few days.
Regulatory application of QSAR and read-accross models
Can computational models be applied at regulatory level?
Chemoinformatic models are recognized as a valid alternative at a regulatory level, and the international regulations themselves encourage increasingly their use. For example, in the case of REACH, ECHA systematically requests companies to demonstrate that they have fully considered the use of alternative methods before concluding that a new test with vertebrates is necessary.
Can computer models replace the experimental tests required by legislation (e.g. REACH)?
Yes, in a lot of cases our models make unnecessary to perform additional assays. This is particularly common when there is already some bibliographic information on the studied compounds, or when the products are of natural origin. In other cases the computational studies do not completely replace the experimental tests, but it is usually possible to reduce them (and therefore reduce time and costs). Another factor to consider is the credibility of the company in front of the authorities, which is clearly favored when presenting a complete dossier including the evaluation by alternative methods.
What type of model do I need for the registration/authorization of my substance: QSAR or read-across?
The type of model depends on the baseline information and on the substance to be evaluated. In principle, QSAR models are technically more robust and reliable, but they cannot always be used (for example, if the compound to be evaluated does not fall within the domain of applicability of the compounds that have been used to generate the models). In that case read-across is a good alternative, since this technique has the same validity at regulatory level when properly executed by an experienced professional.
What is the “Applicability Domain”? How does it affect the prediction of my substance at regulatory level?
Applicability domain is the “chemical space” to which a specific compound belongs in relation to the compounds that have served as the basis for developing a QSAR model. One of the parameters required by the OECD (and therefore necessary from a regulatory point of view) when determining whether a QSAR is valid on a given compound is the characterization of that space.
How are QSAR models validated?
All our models are subjected to a double validation, internal and external, and when possible we perform an experimental validation with independent structures belonging to the same chemical space or “domain of applicability”.