Introduction
At the present, the most common methods for pharmacological compound design include the use of physical-chemical descriptors from QSAR methology, along with the possible complementary addition of quantum mechanics calculations or graphic methods based on molecular mechanics.
An alternative method, based on molecular topology and called «Molecular Connectivity», consists on numerically characterizing the molecule in study by a series of indexes that are specific and exclu-sive for each one.
The aim of the method is to obtain multi-lineal correlation between physical, chemical and biologi-cal properties of molecules, after their topology quantification. For this, correlation functions are ob-tained between these properties (connectivity functions) and a series of descriptors called topological indexes.
This technique has been applied to a group of diverse anti-neoplasic compounds finding connec-tivity functions that are capable of discriminating if a particular compound has cytotoxic activity or not.
Methods and Calculations
In this work, 62 indexes were used for determination of connectivity functions. Hence, regression functions that describe each property were obtained by correlation of experimental values of properties with use of statistic packages for multi-lineal correlation.
In order to classify the chemicals by their anti-neoplasic activity, an equation was defined by use of discriminating lineal analysis and working on a database of about 12 thousand compounds. A large group of chemicals were selected and distributed into two subgroups: one with contrasted anti-neoplasic activity and another for which this activity has not been yet described.
Using connectivity indexes, correlation functions were chosen for different properties and used as filters for selection of possible anti-neoplasics. From application of the discriminating equation of choice, two pharmacological compounds, for which no anti-neoplasic activity had been described, were chosen.
Results and Discussion
The chosen discriminating function was:
D= - 9.06457 - 1.5237 2 XV + 2.06966 4Xp – 18.54615 J2 + 34.43409 J2V
Once applied to the selected group of compounds, it correctly classified 90% of the active and 93.1% of the inactive compounds. The use of this method allows finding new active compounds within series defined by particular structural conditions.
In the present work, two potential cytostatic compounds were selected:
The first of these compounds has an activity defined as anti-spasmodic and coronary vasodilator, while tricromil is known as an analgesic and anti-inflammatory.