Generalized linear models and mixed linear models can take into account confounding factors (age, gender, others...) when analyzing the association between phenotypes and genoypes. Besides, MLM can take into account the "clusters" in the population studied. For instance, it can take into account the family in a human population. If you work on animals models, such Drosophila, you can study 200 lines, 10 individuals/line with the same genetic background, and MLM takes this into account in the statistic model.