Project description:The Caucasian lynx, Lynx lynx dinniki, has one of the southernmost distributions in the Eurasian lynx range, covering Anatolian Turkey, the Caucasus and Iran. Little is known about the biology and the genetic status of this subspecies. To collect baseline genetic, ecological and behavioural data and benefit future conservation of L. l. dinniki, we monitored 11 lynx territories (396 km2) in northwestern Anatolia. We assessed genetic diversity of this population by non-invasively collecting 171 faecal samples and trapped and sampled 12 lynx individuals using box traps. We observed high allelic variation at 11 nuclear microsatellite markers, and found no signs of inbreeding despite the potential isolation of this population. We obtained similar numbers of distinct genotypes from the two sampling sources. Our results indicated that first order female relatives occupy neighbouring territories (female philopatry) and that territorial male lynx were highly unrelated to each other and to female territorial lynx, suggesting long distance male dispersal. Particular male and female resident territorial lynx and their offspring (kittens and subadults) were more likely to be trapped than resident floaters or dispersing (unrelated) lynx. Conversely, we obtained more data for unrelated lynx and higher numbers of territorials using non-invasive sampling (faeces). When invasive and non-invasive samples were analysed separately, the spatial organisation of lynx (in terms of female philopatry and females and males occupying permanent ranges) affected measures of genetic diversity in such a way that estimates of genetic diversity were reduced if only invasive samples were considered. It appears that, at small spatial scales, invasive sampling using box traps may underestimate the genetic diversity in carnivores with permanent ranges and philopatry such as the Eurasian lynx. As non-invasive sampling can also provide additional data on diet and spatial organisation, we advocate the use of such samples for conservation genetic studies of vulnerable, endangered or data deficient territorial species.

Project description:The Amazon biome is home to the largest tropical forest on the planet and has the greatest global biodiversity on Earth. Despite this, several less charismatic taxonomic groups, such as amphibians, lack comprehensive studies on their species richness and spatial distribution in the Amazon Region. In this study, we investigated: i) patterns of richness and endemism of Amazonian amphibians across geopolitical and biogeographic divisions, ii) similarities between different Amazonian bioregions, iii) temporal trends in amphibian sampling, iv) conservation status of amphibians according to assessments of the IUCN and v) the importance of diverse data sources in building a robust database of amphibian occurrences. We aggregated data from four different sources: publicly accessible platforms, peer-reviewed articles, grey literature and fieldwork inventories spanning 15 years (2007-2021), ultimately compiling 160,643 records of 947 species across 7,418 sampled sites. The greatest diversity of species was found in Peru, Brazil and Ecuador, with notable amphibian diversity and endemism in regions such as the western basins and the Tapajós River Basin in the central-southern Amazon. Geographical analysis of species diversity revealed four distinct groups defined by latitudinal (the Amazon River) and longitudinal (the Juruá, Madeira and Tapajós Rivers) gradients, with low species similarity (< 40%), particularly in the basins of north-western Amazonia. Amphibian sampling in the Amazon has intensified since the 1950s with the establishment of important research centres such as INPA and the GOELD Museum in the Brazilian Amazon. Approximately 18% of Amazonian amphibian species face extinction risk, according to IUCN assessments, highlighting the need for comprehensive data sources to understand and conserve species in this megadiverse region. Our findings suggest that river systems likely influence Amazonian amphibian species composition due to biogeographic history, emphasising the need for robust taxonomic and spatial databases. This study, therefore, contributes a valuable large-scale dataset for Amazonian amphibians, guiding future research and strategies for amphibian conservation.

Project description:Task-optimized convolutional neural networks (CNNs) show striking similarities to the ventral visual stream. However, human-imperceptible image perturbations can cause a CNN to make incorrect predictions. Here we provide insight into this brittleness by investigating the representations of models that are either robust or not robust to image perturbations. Theory suggests that the robustness of a system to these perturbations could be related to the power law exponent of the eigenspectrum of its set of neural responses, where power law exponents closer to and larger than one would indicate a system that is less susceptible to input perturbations. We show that neural responses in mouse and macaque primary visual cortex (V1) obey the predictions of this theory, where their eigenspectra have power law exponents of at least one. We also find that the eigenspectra of model representations decay slowly relative to those observed in neurophysiology and that robust models have eigenspectra that decay slightly faster and have higher power law exponents than those of non-robust models. The slow decay of the eigenspectra suggests that substantial variance in the model responses is related to the encoding of fine stimulus features. We therefore investigated the spatial frequency tuning of artificial neurons and found that a large proportion of them preferred high spatial frequencies and that robust models had preferred spatial frequency distributions more aligned with the measured spatial frequency distribution of macaque V1 cells. Furthermore, robust models were quantitatively better models of V1 than non-robust models. Our results are consistent with other findings that there is a misalignment between human and machine perception. They also suggest that it may be useful to penalize slow-decaying eigenspectra or to bias models to extract features of lower spatial frequencies during task-optimization in order to improve robustness and V1 neural response predictivity.

Dataset Information

Robustness of genetic diversity measures under spatial sampling and a new frequency-independent measure

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets