Understanding the assembly processes of symbiont communities, including viromes and microbiomes, is important for improving predictions on symbionts’ biogeography and disease ecology. Here, we use phylogenetic, functional and geographic filters to predict the similarity between symbiont communities in viral communities of Mexican bats.

METHODS

We constructed a database using all reports of viruses isolated or detected by molecular techniques from the American continent and associated with Mexican chiropters. The information used was that reported by Rico-Chávez et al, 2015 in their PCR analyses with bat samples obtained from the south of Mexico and information available online at DBatVir (http://www.mgc.ac.cn/DBatVir/).

Two models were tested: a “known” model (Appendix 1), where only reports from Mexico were used, and a “potential” model (Appendix 2), which included virus reports from all of the Americas, but restricted to only those bat species that are present in Mexico, but which have not necessarily been identified as hosts of a given virus there. 

For each model, a virus-host adjacency matrix, cataloged as presence/absence of the virus in each host species, was constructed. The associated Jaccard index (Jaccard, 1922; Real and Vargas, 1996; Cooper et al., 2012) was calculated using the vegan library in free software R (Oksanen et al., 2016; R Core Team, 2017), with a value range between 0, corresponding to completely different viral communities, and 1, corresponding to identical viral communities for each host pair.

For the filters: the spatial filter was represented using a co-occurrence measure, e, a binomial test, measures the statistical significance of the difference between the conditional probability, P(C|X), for two bat species C and X to co-occur, relative to the null hypothesis, P(C). These probabilities are calculated by laying a uniform grid of N cells on the geographic area of interest and then calculating P(C ) = N(C)/N, P(X ) = N(X)/N and P(C|X) = N(CX)/N(X),  where N(C ) and N(X) are the number of cells with presence of the bat species C or X respectively and N(CX) is the number of cells with presence of both, Presence or no presence of the different hosts was evaluated using point collection data from the C3/CONABIO platform SPECIES (http://species.conabio.gob.mx/) using a uniform grid of 6,473 cells of area 20km². In this database, the collections reported by the Mexican research institutes are registered, so there are 128,071 records of chiropterans that have been collected from 1889 to the present along with their sampling site. The denominator in e is the standard deviation of the binomial distribution. In the case where this distribution may be approximated by a normal distribution |e | > 1.96 corresponds to the standard 95% confidence interval that the co-distribution of C and X is not consistent with the null hypothesis that they are distributed independently. The importance of using e as opposed to P(C|X) directly is that the latter can introduce spurious correlations in the case where N(X) is small.

As a phylogenetic filter, the phylogenetic distance between hosts was calculated and the results were normalized using the formula:  log (x) – log (ẋ), where we calculate the logarithm of the phylogenetic distance for a pair of species and subtract the logarithm of the mean phylogenetic distance. This value was extracted from the mammalian super tree (Bininda-Emonds et al., 2007) using the library picante (Kembel et al., 2010) implemented in R (R Core Team, 2017).

For the functional filter, the characteristics evaluated were trophic guild and body mass (Bordes, Morand and Ricardo, 2008; Han et al., 2015; Luis et al., 2015; Rico-Chávez et al., 2015). The information was obtained using PanTHERIA (Jones et al., 2009), Animal Diversity Web (http://animaldiversity.org) and the Encyclopedia of Life (http://eol.org/). The trophic guild was classified as “(1) herbivore (not vertebrate and/or invertebrate), (2) omnivore (vertebrate and/or invertebrate plus any of the other categories) and (3) carnivore (vertebrate and/or invertebrate only)” (Jones et al., 2009). We assigned the value one to a pair of hosts when the trophic guild was the same and zero when it was different. For body mass, we used two metrics, the absolute difference between the pair of bats and the average body mass in grams of the pair.