Song Richness Index: A measure to understand the diversity and repetition of notes in a birdsong

Birdsong plays an important role in mate attraction and territorial defense. Many birds, especially Passerines, produce varying sequences of multiple notes resulting in complex songs. Studying the diversity of notes within these songs can give insights into an individuals reproductive fitness. We first looked at the previously described and commonly used diversity measures to understand the possible case-specific limitations. We then developed a new diversity measure- Song Richness Index (SRI). We compared SRI with three measures of diversity using all possible combinations of notes to understand the case-specific advantages and limitations of all approaches. Simulating all possible combinations gave us insights into how each diversity measure works in a real scenario. SRI showed an advantage over conventional measures of diversity like Note Diversity Index (NDI), Shannons Equitability (SH), and Simpsons Diversity (SI), especially in the cases where songs are made up of only one type of repetitive note.

Characterization of the endometrial, cervicovaginal and anorectal microbiota in post-menopausal women with endometrioid and serous endometrial cancers

Objective To characterize the microbiota of postmenopausal women undergoing hysterectomy for endometrioid (EAC) or uterine serous cancers (USC) compared to controls with non-malignant conditions. Methods Endometrial, cervicovaginal and anorectal microbial swabs were obtained from 35 postmenopausal women (10 controls, 14 EAC and 11 USC) undergoing hysterectomy. Extracted DNA was PCR amplified using barcoded 16S rRNA gene V4 primers. Sequenced libraries were processed using QIIME2. Phyloseq was used to calculate α- and β- diversity measures. Biomarkers associated with case status were identified using ANCOM after adjustment for patient age, race and BMI. PICRUSt was used to identify microbial pathways associated with case status. Results Beta-diversity of microbial communities across each niche was significantly different (R2 = 0.25, p < 0.001). Alpha-diversity of the uterine microbiome was reduced in USC (Chao1, p = 0.004 and Fisher, p = 0.007) compared to EAC. Biomarkers from the three anatomical sites allowed samples to be clustered into two distinct clades that distinguished controls from USC cases (p = 0.042). The USC group was defined by 13 bacterial taxa across the three sites (W-stat>10, FDR<0.05) including depletion of cervicovaginal Lactobacillus and elevation of uterine Pseudomonas. PICRUSTt analysis revealed highly significant differences between the USC-associated clades within the cervicovaginal and uterine microbiota. Conclusions The microbial diversity of anatomic niches in postmenopausal women with EAC and USC is different compared to controls. Multiple bacteria are associated with USC case status including elevated levels of cervicovaginal Lactobacillus, depletion of uterine Pseudomonas, and substantially different functional potentials identified within cervicovaginal and uterine niches.

Effect of chemotherapy and weight change on the gut microbiome of breast cancer patients during the first year of treatment

Objective: The effects of chemotherapy and weight changes on the gut microbiome of breast cancer patients are not well understood. Methods: We conducted a 1-year follow-up study of 33 breast cancer patients and investigated gut microbiome before initiation of chemotherapy and after completion of treatment. We compared alpha diversity and mean taxa abundance at baseline and absolute changes (Δ; final-baseline) in taxa abundance by treatment (16 neoadjuvant- neoADJ, 13 adjuvant- ADJ, 4 no chemotherapy-noC) using Wilcoxon rank sum and negative binomal tests and evalauted whether these changes were affected by weight changes during follow-up. Results: Alpha diversity measures increased in the neoADJ (+16.4% in OTU p =0.03; +51.6% in Chao1 p =0.03; +7.0% in Shannon index P=0.02; +11.0% in PD whole tree p =0.09) but not in the non-neoADJ group (ADJ+noC). The difference in change in Chao1 index between groups was statistically significant ( p NEOADJ vs ADJ+noC =0.04). Wilcoxon p values of 0.03 to 0.003 were observed for five taxa: Bacteroidetes (g _Alistipes) , Firmicutes (g_Clostridium, g_Eubacterium, g_Bilophila) and Preteobacteria g_Haemophilus). In the negative binomial analysis, changes in abundance differed at Bonferroni-adjusted p values ≤0.0007 for four taxa: two Bacteroidetes taxa ( g_Alistipes, f _S247 ) and two Firmicutes taxa (g_Catenibacterium, g_Eubacterium ). The negative binomial results remained largely the same when we adjusted for weight changes. Conclusions and Relevance: This pilot longitudinal study showed changes in alpha diversity measures and abundance of select taxa that appeared to differ by chemotherapy type. Further investigations are needed to confirm these findings and to assess the impact of these microbiome changes on patient outcome.

Phanerozoic parasitism and marine metazoan diversity: dilution versus amplification

Growing evidence suggests that biodiversity mediates parasite prevalence. We have compiled the first global database on occurrences and prevalence of marine parasitism throughout the Phanerozoic and assess the relationship with biodiversity to test if there is support for amplification or dilution of parasitism at the macroevolutionary scale. Median prevalence values by era are 5% for the Paleozoic, 4% for the Mesozoic, and a significant increase to 10% for the Cenozoic. We calculated period-level shareholder quorum sub-sampled (SQS) estimates of mean sampled diversity, three-timer (3T) origination rates, and 3T extinction rates for the most abundant host clades in the Paleobiology Database to compare to both occurrences of parasitism and the more informative parasite prevalence values. Generalized linear models (GLMs) of parasite occurrences and SQS diversity measures support both the amplification (all taxa pooled, crinoids and blastoids, and molluscs) and dilution hypotheses (arthropods, cnidarians, and bivalves). GLMs of prevalence and SQS diversity measures support the amplification hypothesis (all taxa pooled and molluscs). Though likely scale-dependent, parasitism has increased through the Phanerozoic and clear patterns primarily support the amplification of parasitism with biodiversity in the history of life. This article is part of the theme issue ‘Infectious disease macroecology: parasite diversity and dynamics across the globe’.

Using Diversities to Model the Reliability of N-version Machine Learning System

N-version machine learning system (MLS) is an architectural approach to reduce error outputs from a system by redundant configuration using multiple machine learning (ML) modules. Improved system reliability achieved by N-version MLS inherently depends on how diverse ML models are employed and how diverse input data sets are given. However, neither error input spaces of individual ML models nor input data distributions are obtainable in practice, which is a fundamental barrier to understanding the reliability gain by N-version architecture. In this paper, we introduce two diversity measures quantifying the similarities of ML models’ capabilities and the interdependence of input data sets, respectively. The defined measures are used to formulate the reliability of an elemental N-version MLS called dependent double-modules double-inputs MLS. The system is assumed to fail when two ML modules output errors simultaneously for the same classification task. The reliabilities of different architecture options for this MLS are comprehensively analyzed through a compact matrix representation form of the proposed reliability model. Except for limiting cases, we observe that the architecture exploiting two diversities tends to achieve preferable reliability under reasonable assumptions. Intuitive relations between diversity parameters and architecture reliabilities are also demonstrated through numerical experiments with hypothetical settings.

Using Diversities to Model the Reliability of N-version Machine Learning System

N-version machine learning system (MLS) is an architectural approach to reduce error outputs from a system by redundant configuration using multiple machine learning (ML) modules. Improved system reliability achieved by N-version MLS inherently depends on how diverse ML models are employed and how diverse input data sets are given. However, neither error input spaces of individual ML models nor input data distributions are obtainable in practice, which is a fundamental barrier to understanding the reliability gain by N-version architecture. In this paper, we introduce two diversity measures quantifying the similarities of ML models’ capabilities and the interdependence of input data sets, respectively. The defined measures are used to formulate the reliability of an elemental N-version MLS called dependent double-modules double-inputs MLS. The system is assumed to fail when two ML modules output errors simultaneously for the same classification task. The reliabilities of different architecture options for this MLS are comprehensively analyzed through a compact matrix representation form of the proposed reliability model. Except for limiting cases, we observe that the architecture exploiting two diversities tends to achieve preferable reliability under reasonable assumptions. Intuitive relations between diversity parameters and architecture reliabilities are also demonstrated through numerical experiments with hypothetical settings.

Landscape configuration affects probability of apex predator presence and community structure in experimental metacommunities

Biodiversity is declining at an unprecedented rate, highlighting the urgent requirement for well-designed protected areas. Design tactics previously proposed to promote biodiversity include enhancing the number, connectivity, and heterogeneity of reserve patches. However, how the importance of these features changes depending on what the conservation objective is remains poorly understood. Here we use experimental landscapes containing ciliate protozoa to investigate how the number and heterogeneity in size of habitat patches, rates of dispersal between neighbouring patches, and mortality risk of dispersal across the nonhabitat ‘matrix’ interact to affect a number of diversity measures. We show that increasing the number of patches significantly increases γ diversity and reduces the overall number of extinctions, whilst landscapes with heterogeneous patch sizes have significantly higher γ diversity than those with homogeneous patch sizes. Furthermore, the responses of predators depended on their feeding specialism, with generalist predator presence being highest in a single large patch, whilst specialist predator presence was highest in several-small patches with matrix dispersal. Our evidence emphasises the importance of considering how top-down effects can drive community responses to patch configuration.