Search

Abstract As the enthusiasm surrounding Deep Learning grows, both medical practitioners and regulatory bodies are exploring ways to safely introduce image segmentation in clinical practice. One frontier to overcome when translating promising research into the clinical open world is the shift from static to continual learning. Continual learning, the practice of training models throughout their lifecycle, is seeing growing interest but is still in its infancy in healthcare. We present Lifelong nnU-Net, a standardized framework that places continual segmentation at the hands of researchers and clinicians. Built on top of the nnU-Net—widely regarded as the best-performing segmenter for multiple medical applications—and equipped with all necessary modules for training and testing models sequentially, we ensure broad applicability and lower the barrier to evaluating new methods in a continual fashion. Our benchmark results across three medical segmentation use cases and five continual learning methods give a comprehensive outlook on the current state of the field and signify a first reproducible benchmark.

Abstract Studies on the feeding behavior of hematophagous insects, particularly those of medical importance, are relevant for tracking possible pathogen transmission routes and identifying biases in the choice of vertebrates. We evaluated host selection of blood-feeding mosquitoes in a disturbed forest in the Magdalena Medio valley in Colombia from March 2017 to April 2018, after the introduction of Zika virus to the Americas from the 2015–2016 outbreak. We estimated vertebrate diversity and collected blood-engorged female mosquitoes. Genomic DNA/RNA was extracted from the mosquito’s abdomen for vertebrate host identification and pathogen detection. We performed conventional PCR and sequencing, using universal primers targeting vertebrate regions of the eukaryotic mitochondrial genome to determine bloodmeal host. Additionally, we tested for the presence of flaviviruses in all mosquito samples with RT-PCR. Based on the identity and quantity of detected bloodmeals, we performed mosquito-vertebrate interaction network analysis and estimated topology metrics. In total, we collected 292 engorged female mosquitoes representing 20 different species. Bloodmeal analyses identified 26 vertebrate species, the majority of which were mammals (N = 16; 61.5%). No flaviviruses of medical importance were detected from the samples. Although feeding patterns varied, network analyses showed a high degree of specialization by mosquitoes and revealed ecological and phylogenetic relationships among the host community. We conclude that host selection or preference by mosquitoes is species specific.

Abstract Microbiome is known to play an important role in the health of organisms and different factors such as diet have been associated with modifications in microbial communities. Differences in the microbiota composition of wild and captive animals has been evaluated; however, variation during a reintroduction process in primates has never been reported. Our aim was to identify changes in the bacterial composition of three individuals of reintroduced woolly monkeys (Lagothrix lagothricha) and the variables associated with such changes. Fecal samples were collected and the V4 region of the 16S rRNA gene was sequenced to determine gut microbial composition and functionality. Individual samples from released individuals showed a higher microbial diversity after being released compared to before liberation, associated with changes in their diet. Beta diversity and functionality analysis showed separation of samples from released and captive conditions and the major factor of variation was the moment of liberation. This study shows that intestinal microbiota varies depending on site conditions and is mainly associated with diet diversity. The intake of food from wild origin by released primates may promote a positive effect on gut microbiota, improving health, and potentially increasing success in reintroduction processes.

Abstract The Leishmaniases are a group of neglected tropical diseases caused by different species of the protozoan parasite Leishmania, transmitted to its mammalian hosts by the bites of several species of female Phlebotominae sand flies. Many factors have contributed to shifts in the disease distribution and eco epidemiological outcomes, resulting in the emergence of Cutaneous Leishmaniasis outbreaks and the incrimination of vectors in unreported regions. New research development is vital for establishing the new paradigms of the present transmission cycles, hoping to facilitate new control strategies to reduce parasite transmission. Hereafter, this work aims to model and infer the current transmission cycles of Cutaneous Leishmaniasis in Colombia defined by vector and mammal species distributed and interacting in the different regions and validate them by performing sand fly and mammal collections. Vector-host co-occurrences were computed considering five ecoregions of the Colombian territory defined by the World Wide Fund for Nature (WWF) and downloaded from The Nature Conservancy TNC Maps website. Four validation sites were selected based on Cutaneous Leishmaniasis prevalence reports. Sand flies and mammals captured in the field were processed, and species were defined using conventional taxonomic guidelines. Detection of infection by Leishmania was performed to identify transmission cycles in the selected areas. This study uses predictive models based on available information from international gazetteers and fieldwork to confirm sand fly and mammalian species' sustaining Leishmania transmission cycles. Our results show an uneven distribution of mammal samples in Colombia, possibly due to sampling bias, since only two departments contributed 50% of the available samples. Bats were the vertebrates with the highest score values, suggesting substantial spatial overlap with sand flies than the rest of the vertebrates evaluated. Fieldwork allowed identifying three circulating Leishmania species, isolated from three sand fly species. In the Montane Forest ecosystem, one small marsupial, Gracilinanus marica, was found infected with Leishmania panamensis, constituting the first record of this species infected with Leishmania. In the same locality, an infected sand fly, Pintomyia pia, was found. The overall results could support the understanding of the current transmission cycles of Leishmaniasis in Colombia.