Background: Tuberculosis (TB) is a major global public health problem with countries inrnsub-Saharan Africa (SSA) including Ethiopia being most affected by the disease.rnUnderstanding the epidemiology of TB and the development and spread of drug-resistantrnTB is limited by scarce molecular epidemiologic data in SSA including Ethiopia. In terms ofrnepidemiology of TB, control of TB in prisons is an important public health priority. In somernsettings, prisons have been recognized as a reservoir and sites of transmission for TB.rnAdditionally, the spread of TB in rural settings deserves further attention. In rural settings,rnseveral factors may contribute to enhanced transmission of TB including either missedrndiagnosis or delays in diagnosis.rnrnObjective: To describe the molecular epidemiology and drug resistance pattern of TB tornbetter understand the transmission of TB and the development and spread of drug resistancernin different settings of Southern Ethiopia.rn rnMethods: Three studies were conducted, all using a cross-sectional design: a prison study inrnHawassa Prison, a health facility-based study at nine health facilities in and aroundrnShashemene areas of Western Arisi Zone (Oromia region) and a community-based study inrnthe rural areas of Hawassa Zuria Woreda (Sidama Zone, Southern Region). In the prisonrnstudy, incarcerated individuals were screened for TB using a symptom screen; those with arncough of â‰¥2 weeks had spot and morning sputum samples collected for AFB smearrnmicroscopy and molecular diagnostic testing (Xpert MTB/RIF). The study was conductedrnfrom June 15 through July 13, 2015. The health facility study was conducted among sputumrnsmear positive patients with TB visiting nine diagnostic facilities in Southern Ethiopia.rnThree consecutive sputum samples (spot-morning-spot) per patient were examined usingrnAFB smear microscopy with all smear positive specimens also having AFB culturesrnperformed. Mtb isolates had drug susceptibility testing (DST) performed using the indirectrnproportion method and were genotyped with RD9 deletion typing and spoligotyping.rnSpoligotyping International Types (SIT) and sub-lineages (clades) were assigned accordingrnto the SITVITWEB data base. Geographic information system (GIS) was used to map rnsource sites of strains. Whole genome sequencing (WGS) was done on selected Mtb isolates.rnThe study was conducted from June 2015 to May 2016. The population-based survey wasrnconducted in six Kebeles (the lowest administrative units) of Hawassa Zuria Woreda using arncommunity-based active case finding (ACF) strategy. Volunteer women community workersrnled a symptom screening program to identify community members with TB. Individualsrnwith cough â‰¥ 2 weeks were asked to provide two sputum samples (spot and morning) forrndiagnostic testing for TB. AFB smear microscopy and culture and Xpert MTB/RIF werernperformed. The study was conducted from May 08, 2016 through June 09, 2016. rn rnResults: Among 2068 prisoners, 372 (18%) had a positive cough screen. The median age ofrnthese 372 persons was 23 years, 97% were male and 63% were from urban areas. Amongrnthem, 8/372 (2%) had a positive AFB sputum smear microscopy result and 31/372 (8%) hadrna positive Xpert TB/RIF. The point prevalence of pulmonary TB at the prison was 1748 perrn100,000 population as defined by Xpert TB/RIF. In the health facility study, among 250rnnewly diagnosed patients with TB, 154 (52%) were male and 143 (57%) came from ruralrnareas. The prevalence of HIV co-infection was 4%. A total of 230 isolates were obtainedrnfrom 250 sputum cultures. All 230 isolates were M. tuberculosis strains belonging to threernlineages: Euro-American, 187 (81%); East-African-Indian, 31 (14%); and Lineage 7rn(Ethiopian lineage), 8 (4%). The 230 isolates could be categorized into 65 differentrnspoligotype patterns, of which 36 (55.0%) were already known in the international data basernand 29 (45.0%) were new patterns (orphans). The dominant spoligotypes were SIT149rn(21%), SIT53 (19%) and new strains (16%). One hundred ninety-three (84%) isolates werernclustered into 29 spoligotype patterns and the remaining (37, 16%) strains fell into singlernspoligotypes, and clustering of strains by geographic locations was observed. DST revealedrnthat 14% of Mtb isolates tested were resistant to > 1 first line anti-TB drugs and 11% tornisoniazid. SIT 149 was the most prevalent spoligotype among drug resistant isolates. WGSrnanalysis identified different drug resistance mutations in Mtb isolates including onrnpyrazinamide which was not detected by phenotypic DST. MDR-TB was not identified bothrnby phynotipic DST and WGS analysis. For the community study, all 24,517 adults in thernstudy area had a symptom screen performed using 350 health development armies (HDAs);rn544 (2.2%) had cough â‰¥ 2 weeks. Among persons with a positive symptom screen, 13/544 rn(2.4%) had a positive sputum AFB smear microscopy, 13/544 (2.4%) a positive culture andrn32/544 (5.8%) a positive Xpert MTB/RIF test (including two with rifampin resistancernidentified). Overall, 34/544 TB cases (6%) were identified by culture and/or Xpert, whichrncorresponds to a prevalence of 139 per 100,000 population (95% CI: 96-194). rnConclusion: Epidemiological data on TB including drug resistance in different settings isrnvital to look at the problem from different perspectives. A high prevalence of TB wasrndetected among inmates at a large Ethiopian prison. Active case finding using coughrnsymptom screening in combination with the Xpert test has high utility as a potential tool torninterrupt transmission of M. tuberculosis in correctional facilities in high burden, low- andrnmiddle-income countries. The health facility study identified a heterogeneous pool of M.rntuberculosis strains with several clusters including lineage 7 strains circulating in SouthernrnEthiopia. The presence of several clusters and of new strains of M. tuberculosis suggestsrnrecent transmission of TB, including of drug-resistant strains in southern Ethiopia. This callsrnfor regular surveillance of drug susceptibility and wider monitoring and geospatial analysisrnof transmission trends to control TB in the southern Ethiopia. The community studyrndemonstrated the capability of community health workers (volunteer and paid) to rapidlyrnconduct a large-scale population TB screening evaluation and highlighted the high yield ofrnsuch a program to detect previously undiagnosed cases when combined with XpertrnMTB/RIF testing. This could be a model to implement in other similar settings. In general,rnthe results of the three study findings could significantly impact the TB control program ofrnthe country in improving the prevention and control of the disease and add knowledge to thernscience.