Introduction. The research article in English has recently been under scrutiny by theoretical and applied linguists. The understanding of research article is heterogeneous: it can be defined as a scientific manuscript, an independent text type or a separate genre. In Russian academic discourse the concept of the research article usually implies a scientific publication with the definite word count which is less than a monograph. However, this broad definition comprises a variety of heterogeneous genres. The English academic discourse, on the contrary, specifies research genres manifested in Research, Review, Editorial, Commentary, Clinical Case Report and other text names. Additionally, these research genres can vary according to the scientific discipline. The objective of this paper is to identify and classify academic research genres in medical discourse.Methodology and sources. The research corpus was collected from the original highimpact open-access medical journals, i.e., Lancet, New England Journal of Medicine, British Medical Journal, Clinical Infectious Diseases, Journal of Clinical Investigation, Brain, Pediatrics, Diabetes, Heart, Journal of Neuroscience. The research procedure involved contextual, semantic and comparative analysis of the journal requirements on the article type and content presented in typical sections About the Journal, Authors Guidelines and Table of Contents.Results and discussion. The analysis has led to the development of differential parameters for further research genres classification. The findings have shown that a variety of research papers under different names can be classified as a system of research genres in the academic discourse represented by medical research publications. We have also found distinct correlations between medical journal requirements and linguistic characteristics of medical research genres.Conclusion. The academic medical discourse functions in a wide spectrum of article types, which can be classified as medical research genres according to discourse parameters.