In recent years, there has been explosive growth in the amount of biomedical publications. In this paper, we propose a semantic framework that aims to automatically generate an ontology by extracting assertions and topics from multiple free-text scientific publications in PubMed. The pipeline approach for knowledge discovery and ontology generation in the proposed framework has been implemented on the Spark parallel engine based on the Stanford CoreNLP for Natural Language Processing, TF-IDF (Term Frequency Inverse Document Frequency) for feature extraction, OpenIE (Open Information Extraction) for relation extraction, K-Means clustering for topic discovery and OWL API for ontology generation. We have shown that the ontology generated may be very effective in biomedical applications (such as paper search and summarization) with scientific publications.