In this talk, I will present our study on identifying informative sentences in Medline abstracts and classifying semantic relations that exist between diseases and treatments in biomedical sentences. We focus on three semantic relations: Cure, Prevent, and Side Effect. The presentation will describe various representation techniques and classification algorithms that were tried. Based on our results, we suggest that a suitable way to classify disease-treatment semnatic relations in biomedical sentences is to first weed out non-informative sentences and then classify the sentences by the types of relations.