Study sites and sample collection
The study was carried out in Adjumani (3.3784° N, 31.7822° E) and Moyo districts (3.6527° N, 31.7281° E) in northwestern Uganda (Fig. 1). Northwestern Uganda has traditionally been known as an endemic sleeping sickness focus but just like in the rest of the continent, there has been a gradual decline in the number of new cases. For example, passive screening conducted in 2014, detected just nine cases in the region , and in 2018, out of the 10,000 individuals screened, none was infected with T. b. gambiense . In the same year, a tsetse control intervention was expanded to cover the main gambiense HAT foci in West Nile to curtail transmission of gambiense, and this reduced the prevalence to below the elimination threshold (1 new case per 10,000 population), making elimination on course across this study area if status is maintained . Regarding AAT, there are numerous confirmed and anecdotal reports of the presence of the disease among cattle keeping households. Moreover, a recent study  found the prevalence of T. b.brucei in local cattle, pigs and tsetse flies are as follows, 1.9%, 6.3% and 1.8%, respectively. Therefore, AAT still remains a public health challenge in the region.
The region has rainfall periods that run from March-May and from July-November, with a short dry spell in June and a fairly long period of dryness from December to February. The vegetation is a mixture of forests and savannah, with open woodland, grassland, and shrubs.. Besides, there are several fast running streams passing through subsistence farms with low plains and rolling hills and valleys that slope towards the river Nile. The typical riverine habitats are suitable for Glossina fuscipes fuscipes, the principal tsetse vector in this area. The population is largely rural, practicing mixed crop and livestock farming, consisting of food and cash crops such as tobacco, and livestock, mainly cattle, goats, sheep and pigs.
Tsetse trapping was conducted using biconical traps  baited with acetone and cow urine, deployed along suitable habitats, targeting majorly areas of human and animal activities (Fig. 2). At each site, an average of 10 traps were mounted approximately 100 m apart in different vegetation types for three consecutive days and in well shaded areas to minimize fly mortality due to excessive heat. The traps were deployed in the villages of Olwi, Olobo, Oringya, Osugo East and Pagirinya in Adjumani district, and in Orubakulem, Ori, Lea, Cefo and Moyipi villages in Moyo district. The geographical coordinate of each trap was recorded using a handheld Global Positioning Systems (GPS) unit (GPS 12 XL, Garmin Ltd. 2003, Olathe, Kansas, USA). We also recorded the trap codes, sex of the captured flies and dates of collection. To prevent the attack of ants on the flies in the traps, each supporting pole was smeared with grease.
Trapped tsetse flies were collected every 24 hours for at least three consecutive days  . After each collection, tsetse flies were identified morphologically, counted and sorted into teneral and non-teneral as described by Laveissière et al. . The tsetse flies were assigned to one of the six age categories, according to the degree of wear or fraying observed on the hind margin of the wing as described by Jackson . After categorizing the wing fray (WF), the actual age of the fly was estimated using directions for estimating the mean age of a sample of tsetse flies as outlined in the FAO Training manual for tsetse control personnel . The ages of tsetse flies based on wing fray categories were later pooled as “young tsetse” (WF1-2), “old tsetse” (WF 3-4) and “very old tsetse” (WF 5-6) for statistical analysis. Only fresh and non-teneral flies were then selected for dissection to obtain midguts, salivary glands and proboscis. The teneral flies were excluded from the analyses. The dissections were carried out as described in the FAO Training Manual for Tsetse Control Personnel . Trapped flies were dissected in phosphate buffered saline in 2% glucose under a dissecting microscope. Samples of midguts, salivary glands and proboscis from dissected flies were preserved in 70% ethanol in sealed eppendorf tubes until required for subsequent DNA extraction and PCR assays.
The ethanol preserved tissues (miguts, proboscis and salivary glands) of each tsetse sample was pooled together in a single tube and genomic DNA extracted using PurelinkTM extraction kit from Invitrogen following the manufacturer’s instructions. The supernatant was used either directly for PCR or stored at -20°C. Prior to their use or storage, DNA samples were electrophoresed in a 1% agarose gel in 0.5× TBE buffer at 100 V for 45 mins. The quality of DNA in the sample was then estimated by comparing florescent yield of the sample with standard cut Lambda DNA run alongside the DNA samples.
Identification of different trypanosome species
To detect trypanosome DNA, we employed the nested PCR protocol described by Cox et al., , using the same primer sequences but with slight modifications in amplification conditions. The outer primer sequences were ITS1 (5-GAT TAC GTC CCT GCCATT TG-3), and ITS2 (5-TTG TTC GCT ATC GGTCTT CC-3) and inner primer sequences ITS3 (5-GGA AGC AAA AGT CGT AACAAG G-3), and ITS4 (5-TGT TTT CTT TTC CTCCGC TG-3). PCR amplifications were performed in two rounds. The first round was performed in a final reaction volume of 20 µL containing 10pmol of each primer, the BioneerAccuPower® PCR premix (Bioneer Corporation), and 2μL of each DNA template. The amplification conditions began with 1 cycle of denaturation at 95 °C for 5min followed by 40 amplification cycles at 94 °C for 1min, 55 °C for 1min, and 72 °C for 2minutes. In the second round, 2μL of the amplified product from the first round was placed in a fresh tube and 20μL of the reaction mixture was added as described above for the outer primers, except that the outer primers (ITS1 and 2) were substituted with the inner primers (ITS3 and 4). The amplifications conditions were identical to the one described for the first PCR round. To minimize bias due to false positives during repeated PCRs, negative controls in which DNA templates were replaced with sterile distilled water as well as positive control DNAs (of each trypanosome species) were included in all PCR reactions. All reactions were carried out using a GeneAmp 9700 thermal cycler PCR system (Applied Biosystems). After the nested PCR, 5μL of the amplified products were loaded on a casted 1.8% agarose gel, which was subsequently stained with a Gel Red nucleic acid stain, with a 75bp gene marker. These were run in a Mupid®-exu Sub-marine electrophoresis gel tank (Helix Technologies Inc. MEXO 0800137) for 45 mins at 100V in 0.5X TBE buffer. The gels were then visualized under ultra-violet illumination and photographed. Trypanosomes species and subspecies were identified by comparing the molecular sizes of their DNA fragments with the documented band sizes of trypanosome species  (Table 1). For T. brucei, further investigation was done by running a second PCR for diagnosis of T. b. gambiense employing a nested-PCR with a first reaction using TgsGP1/2 primers  and a second one with TgsGP sense2/antisense2 primers described by Morrison et al. .
Identification of tsetse blood meal sources
To test for the origin of blood meals, samples of DNA from residual blood meal in tsetse midguts were amplified using PCR with universal primers complementary to the conserved region of mitochondrial DNA (mtDNA) CO1 and cytb gene as detailed by Muturi et al., . The primer sequences for the CO1 gene was VF1d_t1 (5’ TGTAAAACGACGGCCAGTTCTCAACCAACCACAARGAYATYGG- 3’) and VR1d_t1 (5’-CAGGAAACAGCTATGACTAGACTTCTGGGTGGCCRAARAAYCA- 3’) and for the Cyt b was (Cb5’-CCATCCAACATCTCAGCATGATGAAA-3’) and (Cb2 5’-CCCCTCAGAATGATATTTGTCCTCA-3’) as described by Ivanova et al.,  and Kocher et al., , respectively. Both PCR reactions were performed in a total volume of 20µl containing 10pmol of each primer, 10 mMTris-Cl, pH 8.3 and 50 mMKCl, 1.5 mM MgCl2, 2.5 mMdNTPs, 2µL of the DNA template and 1unit DreamTaq™ DNA polymerase (Fermentas Life Sciences). The PCR was then carried out in a GeneAmpPCR System 9700 (Applied Biosystems) thermocycler. The conditions for CO1 PCR were as follows: initial denaturation at 94°C for 5 minutes, followed by 45 cycles of denaturation at 94°C for 30 s, annealing at 55°C for 1 minute, and primer extension at 72°C for 30 seconds. For Cyt b PCR, conditions were: initial denaturation at 94°C for 5 minutes, followed by 40 cycles of denaturation at 94°C for 45 s, annealing at 55°C for 45 sec, and primer extension at 72°C for 30 s. Positive and negative controls were included in each PCR assay. The positive controls were cattle genomic DNA obtained from the Animal Health Unit at ILRI Research Laboratories. Negative controls consisted of the Master Mix for each PCR and fresh Milli Q water obtained from the Central Core of the BeCA hub laboratories. To analyze the amplicons, 5 μL of the PCR product, as well as negative and positive controls, was resolved in a 1.8% agarose gel at 80 volts for 45 minutes and visualised under UV transilluminator following gel red staining. The amplified products were visualized under ultra-violet illumination and a picture of the gel taken using the gel documentation and analysis systems mounted with a high-performance CCD Camera-COHU.
The PCR products were gel-purified using the GeneJet™ kit (catalog no K0702 EU) following the manufacturer’s instructions and submitted to the BeCA-ILRI Hub Sequencing Unit (Segolip) for sequencing using Bigdye™ Terminator. Sequencing was done bi-directionally using the inner amplification primers for the CO1 and cyt b genes. Consensus sequences were then generated by contiguation functions of the CLC Main Workbench Version 6.6.2 and manually edited, where necessary, by reference to the chromatograms. Vertebrate species were confirmed by sequence alignments with those already deposited in GenBank database using the Basic Local Alignment Search Tool (BLAST, https://blast.ncbi.nlm.nih.gov/Blast.cgi) . Sequences of a given pair-wise alignment from positive PCR products with high percentage similarity (identity matches of 90-98%) and lowest E- values were selected as the most likely species of host.
The average Apparent density (AD), expressed as the average number of flies caught per trap per day (flies/ trap/ day or FTD), was calculated to obtain the data on tsetse distribution in the area for each trapping site using the formula: FTD=ΣF/T×D, where, ΣF is the total number of tsetse flies caught, T is the number of traps deployed and D is the number of days of trapping in the field. Tsetse infection rates were calculated by dividing the number of flies infected with trypanosomes by the total number of flies analysed, and expressed as percentages. Pearson chi-square goodness of-fit-tests (χ2) were employed to determine the association of tsetse infection rates with fly’s sex, district of origin, and age based on wing fray categories. Values of p-value < 0.05 were considered significant at 95% confidence interval (CI). Independent samples t-test was conducted to compare whether the average number of flies caught per trap per day differed between the two districts under study. All statistical tests were performed using SPSS software (version 21.0.1, SPSS Inc., Chicago, IL, USA).