bioRxiv2026
BackgroundThe ecological dynamics between Trypanosoma parasites and their wild mammalian hosts, such as bats and armadillos, are complex. Recent 18S rRNA metabarcoding studies have reported extraordinary levels of hidden parasite diversity and frequent multi-lineage coinfections within individual wild hosts. However, the boundary between genuine biological coinfection and methodological artifact remains difficult to establish. Based on Gauses principle of competitive exclusion, the mammalian bloodstream represents a highly constrained niche where stable coexistence of identical ecological competitors is theoretically rare. We hypothesize that previously reported hyper-diverse Trypanosoma coinfections are largely bioinformatic artifacts, and that true intra-host dynamics instead favor single-lineage dominance. MethodsTo test this hypothesis, we sequenced samples from 27 wild armadillos (Dasypus novemcinctus) and 26 bats from Ecuador. The 18S rRNA gene was amplified via nested PCR and sequenced using an Oxford Nanopore Technologies MinION platform. We developed a progressively stringent bioinformatics pipeline to evaluate coinfection hypotheses. Raw reads were processed through three alignment scenarios: Lenient, Moderate, and Conservative. These scenarios modulate sequence identity, mapping quality (MAPQ), and coverage thresholds to effectively isolate true biological signals from alignment ambiguity. ResultsUnder lenient alignment parameters, the resulting profiles mirrored previous literature, exhibiting massive apparent intra-host multi-lineage diversity. However, as bioinformatic stringency increased to conservative thresholds ([≥] 98% sequence identity, [≥] 99% coverage, and MAPQ [≥] 30), artifactual pseudo-coinfections collapsed. The highly restricted dataset demonstrated overwhelming single-lineage dominance, validating only three active mixed infections out of the retained samples. Furthermore, our rigorous pipeline isolated rare but genuine biological signals, including the detection of Trypanosoma cruzi marinkellei--historically considered a bat-restricted subgenus--within the terrestrial armadillo cohort. We also confirmed the presence of T. cruzi DTU III (TcIII) in Ecuadorian armadillos, representing a significant biogeographical record for the region. ConclusionsOnce methodological noise is computationally stripped away, active multi-strain Trypanosoma coinfections in the host bloodstream are revealed to be ecologically anomalous. Our findings strongly support the principle of competitive exclusion, suggesting established lineages actively suppress competitors. While Oxford Nanopore sequencing offers necessary resolution for wildlife parasitology, fine-tuning algorithmic parameters is critical to accurately represent host-parasite networks and prevent the artificial inflation of intra-host diversity metrics. Author summaryPrevious studies using DNA metabarcoding have reported that wild mammals, such as bats, frequently harbor complex communities of multiple Trypanosoma parasite lineages simultaneously. However, ecological principles suggest that identical competitors struggle to coexist stably within a constrained environment like the host bloodstream. To investigate whether these reported high coinfection rates reflect true biology or methodological artifacts, we sequenced the 18S rRNA gene of Trypanosoma from 26 bats and 27 armadillos in Ecuador. We processed the sequencing data through computational pipelines with progressively stricter filtering parameters. We observed that under lenient filtering, animals appeared to have highly diverse, mixed infections. Conversely, when strict parameters were applied to remove potential analytical noise, the artificial complexity collapsed, revealing that the vast majority of hosts were dominated by a single parasite lineage. We confirmed only three active mixed infections in our highly restricted dataset. Our findings indicate that active multi-strain Trypanosoma coinfections are rare, aligning with the principle of competitive exclusion. These results highlight the necessity of applying rigorous bioinformatic filters to accurately evaluate host-parasite interactions and avoid overestimating diversity metrics.