Mixture models as a method for comparative sociality: social networks and demographic change in resident killer whales

In studies of social behaviour, social bonds are usually inferred from rates of interaction or association. This approach has revealed many important insights into the proximate formation and ultimate function of animal social structures. However, it remains challenging to compare social structure between systems or time-points because extrinsic factors, such as sampling methodology, can also influence the observed rate of association. As a consequence of these methodological challenges, it is difficult to analyse how patterns of social association change with demographic processes, such as the death of key social partners. Here we develop and illustrate the use of binomial mixture models to quantitatively compare patterns of social association between networks. We then use this method to investigate how patterns of social preferences in killer whales respond to demographic change. Resident killer whales are bisexually philopatric, and both sexes stay in close association with their mother in adulthood. We show that mothers and daughters show reduced social association after the birth of the daughter’s first offspring, but not after the birth of an offspring to the mother. We also show that whales whose mother is dead associate more with their opposite sex siblings and with their grandmother than whales whose mother is alive. Our work demonstrates the utility of using mixture models to compare social preferences between networks and between species. We also highlight other potential uses of this method such as to identify strong social bonds in animal populations. Significance statement Comparing patters of social associations between systems, or between the same systems at different times, is challenging due to the confounding effects of sampling and methodological differences. Here we present a method to allow social associations to be robustly classified and then compared between networks using binomial mixture models. We illustrate this method by showing how killer whales change their patterns of social association in response to the birth of calves and the death of their mother. We show that after the birth of her calf, females associate less with their mother. We also show that whales’ whose mother is dead associate more with their opposite sex siblings and grandmothers than whales’ whose mother is alive. This clearly demonstrates how this method can be used to examine fine scale temporal processes in animal social systems.

Reintroducing species when threats still exist: assessing the suitability of contemporary landscapes for island endemics

Reintroducing species into landscapes with persistent threats is a conservation challenge. Although historic threats may not be eliminated, they should be understood in the context of contemporary landscapes. Regenerating landscapes often contain newly emergent habitat, creating opportunities for reintroductions. The Endangered St Croix ground lizard Pholidoscelis polops was extirpated from the main island of St Croix, U.S. Virgin Islands, as a result of habitat conversion to agriculture and predation by the small Indian mongoose Herpestes auropunctatus. The species survived on two small cays and was later translocated to two islands. Since the 1950s, new land-cover types have emerged on St Croix, creating a matrix of suitable habitat throughout the island. Here we examined whether the new habitat is sufficient for a successful reintroduction of the St Croix ground lizard, utilizing three complementary approaches. Firstly, we compared a map from 1750 to the current landscape of St Croix and found statistical similarity of land-cover types. Secondly, we determined habitat suitability based on a binomial mixture population model developed as part of the programme monitoring the largest extant population of the St Croix ground lizard. We estimated the habitat to be sufficient for > 142,000 lizards to inhabit St Croix. Thirdly, we prioritized potential reintroduction sites and planned for reintroductions to take place during 2020–2023. Our case study demonstrates how changing landscapes alter the spatial configuration of threats to species, which can create opportunities for reintroduction. Presuming that areas of degraded habitat may never again be habitable could fail to consider how regenerating landscapes can support species recovery. When contemporary landscapes are taken into account, opportunities for reintroducing threatened species can emerge.

Normalization by distributional resampling of high throughput single-cell RNA-sequencing data

MotivationNormalization to remove technical or experimental artifacts is critical in the analysis of single-cell RNA-sequencing experiments, even those for which unique molecular identifiers (UMIs) are available. The majority of methods for normalizing single-cell RNA-sequencing data adjust average expression in sequencing depth, but allow the variance and other properties of the gene-specific expression distribution to be non-constant in depth, which often results in reduced power and increased false discoveries in downstream analyses. This problem is exacerbated by the high proportion of zeros present in most datasets.ResultsTo address this, we present Dino, a normalization method based on a flexible negative-binomial mixture model of gene expression. As demonstrated in both simulated and case study datasets, by normalizing the entire gene expression distribution, Dino is robust to shallow sequencing depth, sample heterogeneity, and varying zero proportions, leading to improved performance in downstream analyses in a number of settings.Availability and implementationThe R package, Dino, is available on GitHub at https://github.com/JBrownBiostat/[email protected], [email protected]

Inference of ploidy by leveraging read depth from amplicon sequencing

Variation in ploidy occurs naturally in select plant and animal species. Ploidy variation can also occur spontaneously or be induced during artificial propagation of fish and shellfish. Studying species and systems that have variable ploidy requires techniques to infer ploidy of individuals. Massively parallel sequencing of biallelic SNPs has been used to infer ploidy, but existing techniques have several drawbacks. These include being limited to only comparing a fixed number of ploidies (diploidy, triploidy, and tetraploidy) and requiring that heterozygous genotypes in an individual be identified prior to ploidy inference. We describe a method of inferring ploidy from sequencing of biallelic SNPs based on beta-binomial mixture models. This method is generalized to apply to any ploidy and does not require prior identification of heterozygous genotypes. We demonstrate efficacy of this method for comparing ancestral octoploidy, decaploidy, and dodecaploidy (tetraploidy, pentaploidy, and hexaploidy for the sequenced SNPs) in white sturgeon and diploidy and triploidy in Chinook salmon with amplicon sequencing (GT-seq) data. Results indicated that ploidy could be reliably estimated for individuals based on distinct distribution of log-likelihood ratios (LLR) for known ploidy samples of both species that were tested. Confidence in ploidy estimates increased with sequencing depth. We encourage users to explore the sequencing depths and LLR critical values that provide reliable estimates of ploidy for a given organism and set of SNPs. We expect that the R package provided will empower studies of genetic variation and inheritance in organisms that vary in ploidy naturally or as a result of artificial propagation practices.

Effects of cell cycle variability on lineage and population measurements of messenger RNA abundance

Many models of gene expression do not explicitly incorporate a cell cycle description. Here, we derive a theory describing how messenger RNA (mRNA) fluctuations for constitutive and bursty gene expression are influenced by stochasticity in the duration of the cell cycle and the timing of DNA replication. Analytical expressions for the moments show that omitting cell cycle duration introduces an error in the predicted mean number of mRNAs that is a monotonically decreasing function of η , which is proportional to the ratio of the mean cell cycle duration and the mRNA lifetime. By contrast, the error in the variance of the mRNA distribution is highest for intermediate values of η consistent with genome-wide measurements in many organisms. Using eukaryotic cell data, we estimate the errors in the mean and variance to be at most 3% and 25%, respectively. Furthermore, we derive an accurate negative binomial mixture approximation to the mRNA distribution. This indicates that stochasticity in the cell cycle can introduce fluctuations in mRNA numbers that are similar to the effect of bursty transcription. Finally, we show that for real experimental data, disregarding cell cycle stochasticity can introduce errors in the inference of transcription rates larger than 10%.

Towards automatic classification of pathological epileptic tissue with interictal high frequency oscillations

High frequency oscillations (HFOs) recorded by intracranial electrodes have generated excitement for their potential to help localize epileptic tissue for surgical resection (Frauscher et al., 2017). However, previous research has shown that the number of HFOs per minute (i.e. the HFO “rate”) is not stable over the duration of intracranial recordings. The rate of HFOs increases during periods of slow-wave sleep (von Ellenrieder et al., 2017), and HFOs that are predictive of epileptic tissue may occur in oscillatory patterns (Motoi et al., 2018). We sought to further understand how between-seizure (i.e. “interictal”) HFO dynamics predict the seizure onset zone (SOZ). Using long-term intracranial EEG from 16 subjects, we fit Poisson and Negative Binomial mixture models that describe HFO dynamics and include the ability to switch between two discrete brain states. Oscillatory dynamics of HFO occurrences were found to be predictive of SOZ and were more consistently predictive than HFO rate. Using concurrent scalp-EEG in two patients, we show that the model-found brain states corresponded to (1) non-REM (NREM) sleep and (2) awake and rapid eye movement (REM) sleep. This work suggests that unsupervised approaches for classification of epileptic tissue without sleep-staging can be developed using mixture modeling of HFO dynamics.

Effects of cell cycle variability on lineage and population measurements of mRNA abundance

Many models of gene expression do not explicitly incorporate a cell cycle description. Here we derive a theory describing how mRNA fluctuations for constitutive and bursty gene expression are influenced by stochasticity in the duration of the cell cycle and the timing of DNA replication. Analytical expressions for the moments show that omitting cell cycle duration introduces an error in the predicted mean number of mRNAs that is a monotonically decreasing function of η, which is proportional to the ratio of the mean cell cycle duration and the mRNA lifetime. By contrast, the error in the variance of the mRNA distribution is highest for intermediate values of η consistent with genome-wide measurements in many organisms. Using eukaryotic cell data, we estimate the errors in the mean and variance to be at most 3% and 25%, respectively. Furthermore, we derive an accurate negative binomial mixture approximation to the mRNA distribution. This indicates that stochasticity in the cell cycle can introduce fluctuations in mRNA numbers that are similar to the effect of bursty transcription. Finally, we show that for real experimental data, disregarding cell cycle stochasticity can introduce errors in the inference of transcription rates larger than 10%.