RESUMEN
Whole genome sequencing (WGS) of Mycobacterium avium complex (MAC) isolates in the clinical laboratory setting allows for rapid and reliable subspecies identification of a closely related complex of human pathogens. We developed a bioinformatics pipeline for accurate subspecies identification and tested 74 clinical MAC isolates from various anatomical sites. We demonstrate that reliable subspecies level identification of these common and clinically significant MAC isolates, including M. avium subsp. hominissuis (most dominant in causing lower respiratory tract infections in our cohort), M. avium subsp. avium, M. intracellulare subsp. intracellulare, and M. intracellulare subsp. chimaera, can be achieved by analysis of only two marker genes (rpoB and groEL/hsp65). We then explored the relationship between these subspecies and anatomical site of infection. Further, we conducted an in silico analysis and showed our algorithm also performed well for M. avium subsp. paratuberculosis but failed to consistently identify M. avium subsp. silvaticum and M. intracellulare subsp. yongonense, likely due to a lack of available reference genome sequences; all the 3 subspecies were not found in our clinical isolates and rarely reported to cause human infections. Accurate MAC subspecies identification may provide the tool and opportunity for better understanding of the disease-subspecies dynamics in MAC infections.