DISCO: A deep learning ensemble for uncertainty-aware segmentation of acoustic signals

Recordings of animal sounds enable a wide range of observational inquiries into animal communication, behavior, and diversity. Automated labeling of sound events in such recordings can improve both throughput and reproducibility of analysis. Here, we describe our software package for labeling elements in recordings of animal sounds, and demonstrate its utility on recordings of beetle courtships and whale songs. The software, DISCO, computes sensible confidence estimates and produces labels with high precision and accuracy. In addition to the core labeling software, it provides a simple tool for labeling training data, and a visual system for analysis of resulting labels. DISCO is open-source and easy to install, it works with standard file formats, and it presents a low barrier of entry to use.

Gene expression vs. sequence divergence: comparative transcriptome sequencing among natural Rhinolophus ferrumequinum populations with different acoustic phenotypes

BACKGROUND

Although the sensory drive hypothesis can explain the geographic variation in echolocation frequencies of some bat species, the molecular mechanisms underlying this phenomenon are still unclear. The three lineages of greater horseshoe bat (Rhinolophus ferrumequinum) in China (northeast, central-east, and southwest) have significant geographic variation in resting frequencies (RF) of echolocation calls. Because their cochleae have an acoustic fovea that is highly sensitive to a narrow range of frequencies, we reported the transcriptomes of cochleae collected from three genetic lineages of R. ferrumequinum, which is an ideal organism for studying geographic variation in echolocation signals, and tried to understand the mechanisms behind this bat phenomenon by analyzing gene expression and sequence variation.

RESULTS

A total of 8190 differentially expressed genes (DEGs) were identified. We identified five modules from all DEGs that were significantly related to RF or forearm length (FL). DEGs in the RF-related modules were significantly enriched in the gene categories involved in neural activity, learning, and response to sound. DEGs in the FL-related modules were significantly enriched in the pathways related to muscle and actin functions. Using 21,945 single nucleotide polymorphisms, we identified 18 candidate unigenes associated with hearing, five of which were differentially expressed among the three populations. Additionally, the gene ERBB4, which regulates diverse cellular processes in the inner ear such as cell proliferation and differentiation, was in the largest module. We also found 49 unigenes that were under positive selection from 4105 one-to-one orthologous gene pairs between the three R. ferrumequinum lineages and three other Chiroptera species.

CONCLUSIONS

The variability of gene expression and sequence divergence at the molecular level might provide evidence that can help elucidate the genetic basis of geographic variation in echolocation signals of greater horseshoe bats.

Purring Crickets: The Evolution of a Novel Sexual Signal

Opportunities to observe contemporary signal change are incredibly rare but critical for understanding how diversity is created and maintained. We discovered a population of the Pacific field cricket (Teleogryllus oceanicus) with a newly evolved song (purring), different from any known cricket. Male crickets use song to attract females from afar and to court females once near. Teleogryllus oceanicus is well known for sexual signal evolution, as exemplified by a recent signal loss. In this study, we characterized the new purring sound and investigated the role of the purr in long-distance and short-distance communication. The purring sound differed from typical ancestral calls in peak frequency, amplitude, and bandwidth. Further, the long-distance purring song facilitated mate location, though the role of courtship purring song is less clear. Our discovery of purring male crickets is an unprecedented opportunity to watch the emergence of a newly evolved sexual signal unfold in real time and has potential to illuminate the mechanisms by which evolutionary novelties arise and coevolve between the sexes.

Geographic vocal variation and perceptual discrimination abilities in male Australian sea lions

Vocal characteristics can vary among and within populations. In species with geographic variation in the structure of vocalizations, individuals may have the ability to discriminate between calls from local and non-local individuals. The ability to distinguish differences in acoustic signals is likely to have a significant influence on the outcome of social interactions between individuals, including potentially mate selection and breeding success. Pinnipeds (seals, fur seals, sea lions and walruses) are highly vocal yet the Australian sea lion (Neophoca cinerea) is the only eared seal known to show geographic vocal variation in male barks. Barks are produced in many social interactions and encode sufficient information for both individual and colony identity to be discriminable. Yet until now, whether males could themselves discriminate these bark differences was unclear. We performed playback experiments in four breeding colonies to investigate whether males can discriminate local from non-local barks. Overall, males responded more strongly to barks from their own colony compared to barks from other colonies regardless of whether those other colonies were close or distant. Competition for females is high in Australian sea lions, but mating periods are asynchronous across colonies. The ability to correctly assess whether a male is from the same colony, thus representing a potential competitor for mates, or merely a visitor from elsewhere, may influence how males interact with others. Given the high cost of fighting, the ability to discern competitors may influence the nature of male-male interactions and ultimately influence how they allocate reproductive effort.

Behavioral Immunity in Insects

Parasites can dramatically reduce the fitness of their hosts, and natural selection should favor defense mechanisms that can protect hosts against disease. Much work has focused on understanding genetic and physiological immunity against parasites, but hosts can also use behaviors to avoid infection, reduce parasite growth or alleviate disease symptoms. It is increasingly recognized that such behaviors are common in insects, providing strong protection against parasites and parasitoids. We review the current evidence for behavioral immunity in insects, present a framework for investigating such behavior, and emphasize that behavioral immunity may act through indirect rather than direct fitness benefits. We also discuss the implications for host-parasite co-evolution, local adaptation, and the evolution of non-behavioral physiological immune systems. Finally, we argue that the study of behavioral immunity in insects has much to offer for investigations in vertebrates, in which this topic has traditionally been studied.

Interspecific variation of calls in clownfishes: degree of similarity in closely related species

BACKGROUND

Clownfishes are colorful coral reef fishes living in groups in association with sea anemones throughout the Indo-Pacific Ocean. Within their small societies, size hierarchy determines which fish have access to reproduction. These fishes are also prolific callers whose aggressive sounds seem to play an important role in the social hierarchy. Agonistic interactions being involved in daily behaviour suggest how acoustic communication might play an important role in clownfish group. Sounds were recorded and compared in fourteen clownfish species (some of which have never been recorded before) to evaluate the potential role of acoustic communication as an evolutionary driving force.

RESULTS

Surprisingly, the relationship between fish size and both dominant frequency and pulse duration is not only species-specific; all the specimens of the 14 species are situated on exactly the same slope, which means the size of any Amphiprion can be predicted by both acoustic features. The number of pulses broadly overlaps among species, whereas the pulse period displays the most variation even if it shows overlap among sympatric species. Sound comparisons between three species (A. akallopisos, A. ocellaris and A. frenatus) having different types of teeth and body shape do not show differences neither in the acoustic waveform nor in the power spectrum.

CONCLUSION

Significant overlap in acoustic features demonstrates that the sound-producing mechanism is highly conservative among species. Differences in the calls of some species are due to size dimorphism and the sound variation might be in this case a by-product. This morphological constraint does not permit a consideration of acoustic communication as the main driving force in the diversification of clownfishes. Moreover, calls are not produced to find mate and consequently are less subject to variations due to partner preference, which restricts the constraints of diversification. Calls are produced to reach and defend the competition to mate access. However, differences in the pulse period between cohabiting species show that, in some case, sounds can help to differentiate the species, to prevent competition between cohabiting species and to promote the diversification of taxa.

Island hopping introduces Polynesian field crickets to novel environments, genetic bottlenecks and rapid evolution

Teleogryllus oceanicus, a cricket native to Australia, was introduced to Hawaii where it encounters a novel natural enemy responsible for their recent rapid evolutionary loss of singing ability. To explore how genetic diversity varies across their broad range, their mode of introduction to Hawaii and nonadaptive influences on the sexual signalling system, we assessed variation at seven microsatellite loci in 19 Australian and island populations. Genetic variability was highest in Australia, intermediate in Oceania and lowest in Hawaii, and differentiation among local populations was a clear function of geographical distance. Hawaiian populations are most closely related to those from the Society Islands and Cook Islands, and a neighbour-joining tree based on D(A) is consistent with movement by Polynesian settlers. We found evidence of bottlenecks in six island populations (including three Hawaiian populations), supporting previous findings in which bottlenecks were implicated in the crickets' loss of singing ability.

Release from bats: genetic distance and sensoribehavioural regression in the Pacific field cricket, Teleogryllus oceanicus

The auditory thresholds of the AN2 interneuron and the behavioural thresholds of the anti-bat flight-steering responses that this cell evokes are less sensitive in female Pacific field crickets that live where bats have never existed (Moorea) compared with individuals subjected to intense levels of bat predation (Australia). In contrast, the sensitivity of the auditory interneuron, ON1 which participates in the processing of both social signals and bat calls, and the thresholds for flight orientation to a model of the calling song of male crickets show few differences between the two populations. Genetic analyses confirm that the two populations are significantly distinct, and we conclude that the absence of bats has caused partial regression in the nervous control of a defensive behaviour in this insect. This study represents the first examination of natural evolutionary regression in the neural basis of a behaviour along a selection gradient within a single species.

Non-visual crypsis: a review of the empirical evidence for camouflage to senses other than vision

I review the evidence that organisms have adaptations that confer difficulty of detection by predators and parasites that seek their targets primarily using sensory systems other than vision. In other words, I will answer the question of whether crypsis is a concept that can usefully be applied to non-visual sensory perception. Probably because vision is such an important sensory system in humans, research in this field is sparse. Thus, at present we have very few examples of chemical camouflage, and even these contain some ambiguity in deciding whether they are best seen as examples of background matching or mimicry. There are many examples of organisms that are adaptively silent at times or in locations when or where predation risk is higher or in response to detection of a predator. By contrast, evidence that the form (rather than use) of vocalizations and other sound-based signals has been influenced by issues of reducing detectability to unintended receivers is suggestive rather than conclusive. There is again suggestive but not completely conclusive evidence for crypsis against electro-sensing predators. Lastly, mechanoreception is highly understudied in this regard, but there are scattered reports that strongly suggest that some species can be thought of as being adapted to be cryptic in this modality. Hence, I conclude that crypsis is a concept that can usefully be applied to senses other than vision, and that this is a field very much worthy of more investigation.

Silent night: adaptive disappearance of a sexual signal in a parasitized population of field crickets

Sexual signals are often critical for mate attraction and reproduction, although their conspicuousness exposes them to parasites and predators. We document the near-disappearance of song, the sexual signal of crickets, and its replacement with a novel silent morph, in a population subject to strong natural selection by a deadly acoustically orienting parasitoid fly. On the Hawaiian Island of Kauai, more than 90% of male field crickets (Teleogryllus oceanicus) shifted in less than 20 generations from a normal-wing morphology to a mutated wing that renders males unable to call (flatwing). Flatwing morphology protects male crickets from the parasitoid, which uses song to find hosts, but poses obstacles for mate attraction, since females also use the males' song to locate mates. Field experiments support the hypothesis that flatwings overcome the difficulty of attracting females without song by acting as 'satellites' to the few remaining callers, showing enhanced phonotaxis to the calling song that increases female encounter rate. Thus, variation in behaviour facilitated establishment of an otherwise maladaptive morphological mutation.

Patterns of acoustic variation in Cicada barbara Stål (Hemiptera, Cicadoidea) from the Iberian Peninsula and Morocco

Field recordings of the calling song and of an amplitude modulated signal produced by males of Cicada barbara from North Africa and the Iberian Peninsula were analysed in order to assess the geographical acoustic variation and the potential usefulness of acoustic data in the discrimination of subspecies and populations. Sound recordings were digitized and the frequency and temporal properties of the calls of each cicada were analysed. In all regions studied, peak frequency, quartiles 25, 50 and 75% and syllable rate showed low coefficients of variation suggesting inherent static properties. All frequency variables were correlated with the latitude, decreasing from south to north. In addition, most acoustic variables of the calling song showed significant differences between regions, and PCA and DFA analyses supported a partitioning within this species between Iberian Peninsula+Ceuta and Morocco, corroborating mtDNA data on the same species. Therefore, the subspecific division of C. barbara into C. barbara barbara from Morocco and C. barbara lusitanica from Portugal, Spain and Ceuta finds support from the present acoustic analyses, a result which is also reinforced by molecular markers.

Rapid evolutionary change in a sexual signal: genetic control of the mutation 'flatwing' that renders male field crickets (Teleogryllus oceanicus) mute

Colonizing events may expose organisms to physical and ecological environments found nowhere else in their range. Novel selection pressures can then influence subsequent rapid evolutionary changes. Here, I investigate the genetics of one such rapid change in the sexual signal of Polynesian field crickets, Teleogryllus oceanicus, that recently colonized the Hawaiian Islands. In Hawaii, T. oceanicus encounter a deadly parasitoid fly found nowhere else in their range. In <20 generations, a wing mutation, flatwing, that eliminates the crickets' song, an important sexual signal, but protects them from the fly, spread to >90% of males on the island of Kauai. I show, using crosses between flatwing males and females from a population that has never contained flatwings, that the song-suppressing mutation is due to a change in a single sex-linked locus. Contemporary evolution of secondary sexual characteristics has only rarely been identified as the result of single-gene changes and never before as a single sex-linked locus, but sex-linked inheritance is thought to facilitate the rapid evolution of these types of traits. Because divergence of sexual signals can influence reproductive isolation, understanding how colonization events and subsequent selection affect signals, and the genetic mechanisms of such change, can shed light on processes likely to play a role in speciation.

Sex differences in immunity in two species of field crickets

Immune defense often differs between the sexes, with males often having a weaker response, at least among many vertebrates. We examined encapsulation ability, a cell-mediated immune response, in laboratory and field populations of two species of field crickets, Teleogryllus oceanicus (Le Guillou, 1841) and Teleogryllus commodus (Walker, 1869), which have different life histories. In the seasonally breeding T. commodus, males show a stronger encapsulation response than females in both the laboratory and the field, although the difference is more marked under field conditions. The aseasonal T. oceanicus showed no sex difference in encapsulation in either field or laboratory samples fed ad libitum, but when food was experimentally reduced, the same pattern of stronger male response emerged. It is possible that this pattern may result from selection on females to increase investment in reproduction when time and energy for breeding are limited, as is more likely for seasonal breeders or animals under food restriction.

The effects of temperature and age on calling song in a field cricket with a complex calling song,Teleogryllus oceanicus(Orthoptera: Gryllidae)

We examined the effects of temperature and age on calling song in the field cricket Teleogryllus oceanicus. Teleogryllus oceanicus has a complex calling song made up of two different kinds of chirp, long and short. The long chirp is made up of three to eight single pulses of sound and the short chirp consists of several paired pulses. The properties of T. oceanicus calling song did not vary with age, but almost every property of the song varied with temperature. Pulse duration, interpulse interval, and pulse rate in both the long and the short chirp varied with temperature. The number of pulses in the long chirp, number of chirps in the short chirp, chirp rate in the short chirp, duration of the long chirp, carrier frequency of both the short and long chirps, and total song duration were also affected by temperature. The duration of the short chirp and the degree of frequency modulation were the only characteristics that did not vary with temperature. Temperature does not affect the properties of the long and short chirps in the same manner. The long chirp decreases in duration with temperature and has fewer pulses, while the short chirp stays the same in duration and contains a higher number of chirps. These data demonstrate that temperature influences calling-song parameters in a field cricket with a complex calling song.

Temporally fluctuating selection of sex-limited signaling traits in the Texas field cricket, Gryllus texensis

Gryllus texensis males produce acoustic mating signals and display extensive heritable variation in when and how much time they spend signaling throughout the night. The goal of this research is to elucidate the potential mechanism responsible for maintaining this heritable variation. Mating signals attract female crickets. In low-density spring populations females select males that signal most often; in high-density fall populations mating appears random with respect to signaling time. Mating signals also inadvertently attract acoustically orienting parasitoid flies; parasitoids are prevalent during the first half of the evening in the fall mating season. I hypothesized that mating signals are influenced by sex-limited temporally fluctuating selection. I predicted how mating signals would respond to this pattern of cyclical selection a priori, and then measured the sexual characters over four successive generations. I provide correlative evidence that mating signals appear to respond to sex-limited temporally fluctuating selection. These results indicate that sex-limited temporally fluctuating selection may play a role in the maintenance of variation in these sexual characters.