Skip to main content

Computerized Molecular Evaluation of Toxicity


We are more and more aware of the need to understand and predict the consequences of the chemicals on the health of human beings and the wild life. Using conventional methods, the assessing of toxic and ecotoxic risks should need funds, time and resources that are not available to solve in detail all the problems related with the increasing number of chemicals. On the other side many data and experiments are available, on thousands of chemicals. Further important information can be obtained from molecular descriptors. Molecular parameters and descriptors have been often used to describe, understand, model and predict the behaviour of chemicals in certain phenomena The explosion of information requires a radical change in the way to deal with the data of information. Artificial intelligence (AI) has been used in a few cases for toxicity prediction and expert systems (ES) in this field have been produced. Today, techniques such as inductive learning algorithms, fuzzy logic, artificial neural networks (ANN) and evolutionary algorithms (EA) can be used to model processes and phenomena with a non-linear behaviour. No studies have been done to compare all the techniques so far introduced using the same set of compounds.
This research will focus on the relationship between chemicals and their toxic and ecotoxic effects investigated through updated computerized approaches. The major objective of the COMET is to extract as much as possible information from the already available knowledge, from toxicity and ecotoxicity databases and suitable molecular descriptors, using advanced computer approaches such as fuzzy logic, ANN, EA, inductive leaming lgorithms and AI. This implies two separate objectives.

A) Previous studies to predict toxicity used or another toxic activity (correlation studies in most of the cases) or molecular descriptors generally QSAR studies). COMET will use both source of knowledge, combining them. In other words, known toxic activities and molecular descriptors will be used to predict an unknown toxicity.
B) Another important objective of COMET is to compare the different computer approaches, to have a better, more systematic knowledge on the performances, capabilities, advantages and disadvantages of the different systems. Since no comparative studies of the results obtained with the same large set of compounds using so many and different approaches have been presented, we think that COMET will significantly contribute to improve the knowledge on some of the specific objectives of the Programme in Areas and Particular attention will be paid to problems of chemical industries interested in toxic and ecotoxic assessment of their products such as pesticides, both in the development of the systems and in their validation. This will be done through a SME that will act as a reference and consultant. It will also evaluate the developed software. Furthermore, another partner, currently involved in a national committee for pollution evaluation, will interact with national regulatory bodies for environmental assessment.

No one has ever attempted to approach toxicity prediction with such a wide spectrum of different innovative computer tools: this can be done only within a multicenter and multidisciplinar consortium. The other important issue of COMET is the fundamental role that the end-users have in leading the systems towards real world problems and applications, since a chemical industry and an institute involved in regulatory bodies for environmental assessment are in the consortium.

Funding Scheme

CSC - Cost-sharing contracts


Istituto di Ricerche Farmacologiche Mario Negri
Via Eritrea 62
20157 Milano

Participants (4)

Piazza Leonardo Da Vinci 32
20133 Milano
Via Sempione 195
20016 Pero
Universitat de Girona - Escuela Politècnica Superior
Plaça Hospital
17071 Girona
Universität - Gesamthochschule Paderborn
59494 Soest