(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



Social network position is a major predictor of ant behavior, microbiota composition, and brain gene expression [1]

['Tomas Kay', 'Department Of Ecology', 'Evolution', 'University Of Lausanne', 'Lausanne', 'Joanito Liberti', 'Department Of Fundamental Microbiology', 'Thomas O. Richardson', 'School Of Biological Sciences', 'University Of Bristol']

Date: 2023-08

The physiology and behavior of social organisms correlate with their social environments. However, because social environments are typically confounded by age and physical environments (i.e., spatial location and associated abiotic factors), these correlations are usually difficult to interpret. For example, associations between an individual’s social environment and its gene expression patterns may result from both factors being driven by age or behavior. Simultaneous measurement of pertinent variables and quantification of the correlations between these variables can indicate whether relationships are direct (and possibly causal) or indirect. Here, we combine demographic and automated behavioral tracking with a multiomic approach to dissect the correlation structure among the social and physical environment, age, behavior, brain gene expression, and microbiota composition in the carpenter ant Camponotus fellah. Variations in physiology and behavior were most strongly correlated with the social environment. Moreover, seemingly strong correlations between brain gene expression and microbiota composition, physical environment, age, and behavior became weak when controlling for the social environment. Consistent with this, a machine learning analysis revealed that from brain gene expression data, an individual’s social environment can be more accurately predicted than any other behavioral metric. These results indicate that social environment is a key regulator of behavior and physiology.

Social insect colonies are highly tractable systems for studying the relationships between organismal biology and the social environment [ 18 ]. They typically show marked division of labor with individuals within the colony specializing in behaviors such as nursing the brood or foraging [ 19 ]. Individuals interact most frequently with other individuals performing the same behavior, leading to behavior-associated community structure in the colony social network [ 16 , 20 ]. Young individuals typically nurse, and with age, they transition to foraging [ 21 – 27 ]. Both behavior and age are associated with brain gene expression [ 28 , 29 ] and microbiota composition [ 30 , 31 ]. Here, we combine automated behavioral tracking with a multiomic approach to simultaneously investigate the correlation structure among social environment, physical environment, behavior, age, brain gene expression, and gut microbiota composition. We used the carpenter ant Camponotus fellah as a model system because the social environment of this species is well characterized and the associations between social environment, age, and foraging behavior have already been quantified [ 16 , 20 ].

In highly social species, physiology and behavior are profoundly and reciprocally intertwined with social environments. Studies in a variety of species have shown intricate links between the social environment and gene expression [ 1 – 3 ], microbiota composition [ 4 ], and behavior [ 5 , 6 ]. However, as is typical in complex biological systems, redundant correlations are ubiquitous: Microbiota composition correlates with behavior [ 4 , 7 – 10 ] and gene expression [ 11 , 12 ], and gene expression is linked to behavior [ 13 , 14 ] and a plethora of other traits. Further, the social environment is often confounded by the physical environment and demographic processes [ 15 – 17 ]. Teasing apart the correlation structure among these variables has therefore been challenging. Moreover, most studies have focused on one or few variables, and the resolution of the social environmental data has been limited.

Results and discussion

We tracked 4 queenright colonies each containing approximately 100 known-age workers (Fig 1). From the automated tracking data, we inferred all pairwise social interactions, determined the spatial distributions of all individuals, and quantified 6 of the most frequent and identifiable task behaviors (tending the queen, foraging, nursing, guarding, trophallaxing, and cleaning). Immediately after behavioral tracking, RNA-sequencing was performed on whole individual brains, and 16S rRNA gene sequencing was performed on surface-sterilized individual abdomens.

C. fellah social networks comprise 2 overlapping communities (groups within which individuals interact frequently and between which individuals interact rarely); one comprising individuals that tend to interact with the queen and brood and the other comprising individuals that tend to leave the nest to forage [20]. Individual position in the social network can be described with “social maturity,” a community detection–based metric that ranges from 0 to 1 and that quantifies the extent to which individuals are associated with the nurse versus the forager social community (see Materials and methods and [20]). Consistent with previous results, social maturity was positively correlated with age and the proportion of time spent foraging (Fig 2; social maturity versus age linear mixed effects regression (LMER) with colony identity as a random factor: R2 = 0.483, t = 20.23, p < 0.001. Social maturity versus proportion of time spent foraging LMER with colony identity as a random factor: R2 = 0.528, t = 22.45, p < 0.001).

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Social network position, time spent foraging, and age. (A) The social networks for each of the 4 colonies (rows) with workers colored according to time spent foraging (column 1), social maturity (column 2), and age (column 3). Lowest values are yellow; highest values are dark blue. Queens are colored magenta. Edge color intensity and width correspond to edge strength. Layouts are calculated with the Fruchterman–Reingold algorithm [32] using R package “iGraph” [33]. (B) Scatter plots relating proportion of time spent foraging, social maturity, and age. The code and data used in this figure are available on Zenodo (doi.org/10.5281/zenodo.8043085 - data: all 4 “Fig 2A…” csv files and “Fig 2B.csv”; code: “02-Main.R”). https://doi.org/10.1371/journal.pbio.3002203.g002

To examine the relationships between brain gene expression profile, microbiota composition, physical environment, social network position, and behavioral profile, we first constructed 5-layer multiplex networks (Fig 3A). In this approach, nodes represent workers and intralayer edges represent pairwise interaction frequency in the social layer, or pairwise similarity (measured by Euclidean distance between profiles) in other layers. Multiple layers show different types of relationships between the same nodes. Inspection of these multiplex networks revealed striking similarities between layers. Individuals with similar behavior also exhibited similar brain gene expression profiles, microbiota compositions, occupied similar physical environments, and social maturities. To compare the strength of the relationships between these 5 layers and age, we reduced each layer to a single dimension (using social maturity for the social layer, and principal component analysis (PCA) for the other layers). We calculated R2 values between these 6 variables and represent the correlations in network form (Fig 3B; values are averages across the 4 colonies; See Fig B in S1 Text for equivalent plots per colony). In this “network-of-networks,” social maturity stands out as a central “hub.” All of the other variables were more correlated with social maturity than with any other variable, except physical environment, which was best correlated with behavior and second best with social maturity. Importantly, both physiological measures (gut microbiota and brain gene expression) were considerably more correlated with social maturity than with behavior, age, or physical environment. The average R2 value between brain gene expression and social maturity was 0.36, 33% greater than the average R2 values between brain gene expression and the physical environment or behavior, and 50% greater than the average R2 value between brain gene expression and age. Similarly, the average R2 value between microbiota composition and social maturity was 0.32, 3% greater than the average R2 value between microbiota composition and behavior, 52% greater than the average R2 value between microbiota composition and physical environment, and 88% greater than the average R2 value between microbiota composition and age.

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Similarity in social network position, physical environment, microbiota, brain gene expression, behavior, and age. (A) A 5-layer multiplex network constructed from behavior, brain gene expression, microbiota, the physical environment, and social interactions. In each layer, each node represents a worker. Nodes are colored according to behavior (cyan = nursing; yellow = cleaning; magenta = foraging; black = guarding). Intralayer edges are unweighted and connect pairs whose interaction strength exceeds the upper quartile of the edge–weight distribution. Interlayer edges connect each worker with itself in the adjacent layers. (B) Graphical representation of the correlation (R2 values) between the 5 layers and age (in blue). Edge width is proportional to edge strength. Layout is calculated with the Fruchterman–Reingold algorithm [32], and vertices are colored according to the layer labels in panel (A) and sized according to their strength (i.e., the sum of their weighted connections). The code and data used in this figure are available on Zenodo (doi.org/10.5281/zenodo.8043085 - data: all 11 “Fig 3A…” txt files and “Fig 3B.csv”; code: “Multiplex.py” and “04-InterlayerCorr.R”). https://doi.org/10.1371/journal.pbio.3002203.g003

The strength of the relationship between brain gene expression and social maturity is of particular interest because it implies that social interactions may have a direct and considerable effect on brain function (i.e., that their association is not an indirect consequence of brain gene expression being associated with behavior or age). Because the PCA of gene expression data could be strongly influenced by few highly expressed genes, we next used differential gene expression analyses to investigate the number of genes differentially expressed by behavior, physical environment, age, microbiota composition, and social maturity. Consistent with the previous analysis, social maturity was associated with the differential expression of the highest number of genes (33% of genes, on average across colonies). Individual behavioral profile was associated with the differential expression of 30% of genes, physical environment with 29% of genes, age with 27% of genes, and microbiota composition with 13% of genes (Table 1). This global pattern was independently true within each colony, meaning that the number of genes differentially expressed as a function of social maturity was significantly higher than the number of genes differentially expressed by behavior in a paired t test (p = 0.024; see Fig C in S1 Text for the percentage of genes differentially expressed by each variable in each colony). This difference became even greater when considering the number of genes differentially expressed by each variable when controlling for each other variable. When controlling for behavior, social maturity still explained the differential expression of 7.3% of genes, whereas when controlling for social maturity, behavior explained the differential expression of only 0.034% of genes. This pattern was also independently true for all 4 colonies (with >10-fold differences in all colonies; see Fig C in S1 Text for full colony-level analyses), strongly supporting the notion that brain transcriptomic variation is more linked to social network position than to behavior.

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Table 1. The number of genes differentially expressed (out of a total of 14,664 genes) in the brain by each of social maturity, behavior, age, the physical environment and gut microbiota composition (column 1), and the percentage of that number that remain differentially expressed when controlling for each other variable (columns 2–6). Few genes remain differentially expressed by behavior, age, the physical environment, or microbiota composition when controlling for the social maturity. https://doi.org/10.1371/journal.pbio.3002203.t001

To further investigate how brain gene expression patterns relate to social maturity, age, and behavior (both overall behavioral profile and the performance of specific tasks), we used a machine learning approach that is more sensitive to nonlinear associations than the above correlational approaches. We iteratively subsampled half of the worker population at random and trained support vector machine models on their gene expression values and the variable of interest. We then used the model to predict the variable of interest from the brain gene expression data for the other half of the worker population and regressed the predicted values against the observed values to quantify predictive accuracy and, hence, the extent to which the variable of interest is reflected in the brain transcriptome. The highest mean R2 between the predicted and observed values (0.76) was obtained for social maturity. The mean R2 values between predicted and observed scores were significantly lower for the 7 other aspects of individual biology analyzed (Fig 4; mean R2 between the predicted and observed position along PC1 of behavioral space = 0.63; age = 0.59; proportion of time spent foraging = 0.53; nursing = 0.33; tending to the queen = 0.23; guarding = 0.05; cleaning = 0.02; t test p-values between social maturity R2 values and all other R2 values all <0.01). These results reinforce the suggestion that there is a fundamental link between social network position and brain gene expression and confirm that this link is stronger than that between task behaviors and brain gene expression.

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. Validation of predictive accuracy. (A) Box plots of the R2 values between observed and predicted values for the proportion of time spent performing each behavior individually, for position along PC1 of behavioral space, for age, and for social maturity. Black lines indicate median values; boxes and whiskers indicate upper and lower quartiles and 1.5× IQ range, respectively. (B) Scatter plot of the predicted versus observed social maturity scores for 10 randomly selected iterations. Color indicates iteration. The code and data used in this figure are available on Zenodo (doi.org/10.5281/zenodo.8043085 - data: “Fig 4A.csv”; code: “05-ML.R”). https://doi.org/10.1371/journal.pbio.3002203.g004

Overall, brain gene expression and microbiota composition correlated more strongly with social network position than with behavior, physical environment, or age. Moreover, while our experiment cannot establish causality or directionality in these relationships, the correlation structure presented here constrains the range of possible causal interactions. If, for example, social interactions were merely a corollary of the spatial distribution of workers (i.e., their physical environments), and if it was physical environment that shaped brain gene expression, then one would expect to see a stronger correlation between the physical environment and brain gene expression than between social environment and brain gene expression. The fact that multiple aspects of behaviorally relevant physiology are more strongly correlated with social interactions than with physical environment, behavior, or age therefore suggests that social interactions may mediate the observed correlations between many aspects of organismal biology and likely play a central role in individual variation in social organisms.

Various factors may limit the generality of these conclusions across species and contexts. First, our experiment was conducted using a eusocial species, and while we would expect the results to hold true for all highly social animals, this remains to be tested. Second, the physical environment was far less complex than those that the ants would naturally experience, which may have reduced the amount of biological variation explained by this variable. Third, composition of the abdominal microbiota of Camponotus is atypical in that it is heavily dominated by Blochmannia, and in some of our workers constituted exclusively Blochmannia. We sequenced to sufficient depth to allow the comparison of the relative abundance of other amplicon sequence variants (ASVs), representing species that were facultatively associated with C. fellah, and observed correlations between microbiota composition and other biological variables that appeared to be mostly driven by the presence of Acetobacteraceae and other species in foragers but not in nurses (Fig G in S1 Text). However, the dominance of Blochmannia and the absence of other bacteria in many individuals may nonetheless reduce the associations between the microbiota composition and the other measured aspects of biology relative to other species.

In conclusion, our study not only highlights the close link between social environments and behavior but also illustrates how social environments relate to behaviorally relevant aspects of physiology, pointing to mechanisms via which individuals can influence each other’s behavior.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3002203

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.