229 / 250 words max.) 6 Human mitochondrial DNA (mtDNA) provides a promising target for microbial source tracking 7 because it is unique to humans and universal across human individuals. We developed a droplet 8 digital PCR (ddPCR) assay, hCYTB484, targeting the cytochrome b gene of the human mtDNA 9 and compared the performance of the hCYTB484 assay with a widely used assay targeting 10 human-associated Bacteroides, the HF183/BacR287 assay. We also defined and validated the 11 analytical limit of detection and analytical lower limit of quantification; these analytical limits 12 determine the concentration levels above which samples are declared to be positive and 13 quantifiable for the target, respectively. We found both assays to be highly specific (95%) 14 against cow and pig feces; however, the hCYTB484 was more sensitive when tested against 15 individual human feces from US, Bangladesh, and Mozambique (100% versus a mean of 51% 16 across the 3 countries). To further compare the performance of the two assays, we utilized a 17 receiver operating characteristic curve, showing that the hCYTB484 marker was widely 18 distributed across human feces from the 3 different geographical regions tested and in higher 19 quantities than the HF183/BacR287 marker. The higher performance of the hCYTB484 marker 20 in individual feces is a desirable characteristic in the detection of fecal pollution from sources to 21 which fewer individuals contribute, such as non-sewered types of sanitation that serve most of 22 Earth's population and carry the highest risk of exposure to fecal-oral pathogens. 23 Importance (142/150 words) 24 The usefulness of an ideal human-specific, culture-and library-independent marker to the 25 microbial source tracking field is reflected by the numerous efforts to develop such markers; 26 however, thus far, microbial-based markers of this type have exhibited variable source-27 specificity across geographies and variable sensitivity across scales of fecal waste. Most of the 28 3 world's population is served by non-sewered forms of sanitation that also carry the highest risk 29 of exposure to fecal-oral pathogens. This reality underscores the need for markers of human fecal 30 contamination that have high sensitivity in fecal pollution sources to which fewer individuals 31 contribute to, such as fecal sludges found in pit latrines. We show that human mtDNA-based 32 methods can be highly source-specific and highly sensitive in smaller scales of fecal pollution, 33 providing a useful addition to the microbial source tracking toolbox by complementing the 34 variable performance of microbial-based markers. 35 4 Word Count 3854/6000 (should not exceed approximately 6,000 words, exclusive of methods, 36 references, figure legends, tables, and supplemental material) 37 Introduction 38Within the microbial source tracking (MST) field, the development of an ideal human-specific, 39 culture-and library-independent marker has been a goal for many research efforts for a variety of 40 reasons-nota...