In the evaluation of particulate matter (PM) toxicity, one faces the challenge of identifying components that could be addressed as markers of toxicity. This study examines the use of statistical methods to determine which components present in the complex mixtures are related to toxic effects. The work is based on data previously published, where we demonstrated that particles collected in different zones of Mexico City showed different ability of inducing cell death, causing DNA damage, or creating proinflammatory effects. Empirically, we correlated those differences to variations in the concentration of transitional metals or endotoxin. In order to test those correlations, in this study we evaluated the role of the concentration of six transitional metals (Ni, V, Zn, Cu, Fe, and Pb) and endotoxin using the following tools: Pearson's regression analysis, correlation matrix between components, and multiple regression followed by a stepwise analysis. Due to collinearity found among several of the components, only Ni, Zn, and Pb were considered as independent variables. Among these, Ni and Zn show better correlations and help to explain the loss in viability, whereas Pb is the component that betters explains the proinflammatory effects. The statistical identification of composition markers facilitates the generation of a hypothesis relating the role played by the composition of PM and its biological effects.