Sex-dependent effects of infection on guppy reproductive fitness and offspring parasite resistance
Data files
Jan 27, 2025 version files 33.59 KB
-
2024TGIPData.csv
28.05 KB
-
README.md
5.54 KB
Abstract
Infection imposes energetic costs on hosts. Hosts typically respond by shifting resources, potentially affecting the quantity and quality of offspring they produce. As the sexes differ in their optimal reproductive strategies, infection of mothers versus fathers may affect offspring quantity and quality in different ways. Here, we test how experimental infection of guppies Poecilia reticulata with the ectoparasite Gyrodactylus turnbulli affects parental reproductive fitness and offspring parasite resistance. We compared breeding pairs in which one or neither parent had previously been infected. In terms of reproductive fitness, parental infection experience did not affect the size, body condition, or number of offspring produced, but fathers who experienced the heaviest infections produced offspring ~55 days sooner than average. This result may represent terminal investment by the males most affected by infection, or may indicate that these males have a faster pace of life, investing in reproduction at the expense of parasite defence. We found that offspring age, parental infection experience, and parental infection severity together strongly predicted offspring parasite resistance. Only among pairs in which one parent had been infected, older offspring, which were those born soonest after the parent’s infection, tended to experience heavier infections. This result may therefore reflect temporary infection-induced reductions in parental investment in offspring quality. Beyond this effect of offspring age, offspring of infected mothers experienced 105 fewer worm days than those of infected fathers: fathers, but not mothers, that experienced heavy infections themselves produced offspring that also experienced heavy infections. The parent-offspring regression for infected fathers is therefore consistent with previous evidence that parasite resistance is heritable in this system, and yields a narrow sense heritability estimate of 0.69±0.13. By contrast, the mother-offspring regression (slope: -0.13±0.17) provides novel insight that mothers may engage in transgenerational immune priming. Overall, our results suggest that the sexes strike a different balance between offspring quantity and quality when faced with infection, with potentially broad implications for disease and host-parasite coevolutionary dynamics in nature.
README: Sex-dependent effects of infection on guppy reproductive fitness and offspring parasite resistance
https://doi.org/10.5061/dryad.7d7wm385d
Description of the data and file structure
Files and variables
File: 20241205_TGIP_supplement.pdf
Description: This .pdf file is an RMarkdown file of the code and output of the analyses we present in the paper.
File: 2024TGIPData.csv
Description: This .csv file contains the raw data from our experiment.
Variables
These variables all refer to the offspring:
- SEX: sex of the fish
- PREL: fish standard length in mm before infection.
- DOB: date in YYYYMMDD the fish was born
- INFTYPE: Whether the fish was infected under anaesthesia through the manual exposure to a heavily infected, euthanized donor fish (‘manual’), or through cohousing with a live infected donor (‘cohouse’)
- LENGTHRESID: The residuals of length of the focal fish on sex. Females tend to be larger than males: using the residuals of this relationship allows us to test for both size and sex differences.
- PREW: Fish weight in mg before infection
- LASTW: Fish weight in mg after infection
- preSMI: Body condition - scaled mass index - before infection
- postSMI: Body condition - scaled mass index - after infection
- deltaSMI: Body condition - scaled mass index - change over the course of infection
- DOSE: Initial number of parasites fish was infected with
- INFBIRTHDAYS: Time in days between the first day of parental infection and fish birth
- LATENCY: Time in days between the day the parental pair was set up to breed and the offspring’s birth, corrected so the minimum value is 0 (as fish were not virgins, some females produced offspring before being paired with a male).
- PAIRBIRTHDAYS: Time in days between the day the parental pair was set up to breed and the offspring’s birth
- MAXWORMS: The maximum number of worms recorded on the fish over its infection
- FULLAUC: The area under the curve of infection load (number of worms) over time - ‘infection integral’
- AGEATINF: Age in days fish was when it was infected
- DATEINF: Date of infection
The datasheet also contains the following variables about the infected parent. Where two variables are listed, those with an ‘M’ prefix refer to the mother, and those with an ‘F’ refer to the father:
- PARINFTREAT: Which parent was infected - mother, father or neither.
- PINF: Binary variable denoting that one parent was (1) or was not (0) infected.
- PAUC: Parental (either the mother’s or father’s, depending on which was infected) area under the curve (AUC) of infection; NA if neither parent infected or parental AUC unknown
- MOTHERID: ID of the parent
- MOTHERINF: Whether (1) or not (0) the parent was infected
- MMAXWORM: The maximum number of worms recorded on the fish over its infection
- MAUC: The area under the curve of infection load (number of worms) over time - ‘infection integral’
- MDATEINF: The date in YYYYMMDD the parent was infected.
- MPREW: Fish weight in mg before infection
- MPOSTW: Fish weight in mg after infection
- MPREL: Fish length in mm before infection
- MPOSTL: Fish length in mm after infection
- MpreSMI: Body condition - scaled mass index - before infection.
- MpostSMI: Body condition - scaled mass index - after infection.
- MdeltaSMI: Body condition - scaled mass index - change over the course of infection.
- MDIED: Whether or not the mother or father died during the observation period
- FATHERID: ID of the parent
- FATHERINF: Whether (1) or not (0) the parent was infected
- FMAXWORM: The maximum number of worms recorded on the fish over its infection
- FAUC: The area under the curve of infection load (number of worms) over time - ‘infection integral’
- FDATEINF: The date in YYYYMMDD the parent was infected.
- FPREW: Fish weight in mg before infection
- FPOSTW: Fish weight in mg after infection
- FPREL: Fish length in mm before infection
- FPOSTL: Fish length in mm after infection
- FpreSMI: Body condition - scaled mass index - before infection.
- FpostSMI: Body condition - scaled mass index - after infection.
- FdeltaSMI: Body condition - scaled mass index - change over the course of infection.
- FDIED: Whether or not the mother or father died during the observation period
- TOTBABIES: Total number of babies produced by that mother
Throughout the datasheet we use "NA" to denote missing values. These values are either missing because this is what is dictated by the experimental design (e.g. some parents were not infected, so will have NAs for all the infection variables), or because fish died before these data could be recorded (e.g. some fish have NA for post infection weights because they died during infection), or because of experimenter error.
Code/software
We used R statistical software to analyse our data. The code and output are included in the pdf provided here. We used the following packages, most up to date versions as of December 2024.
library(glmmTMB) #for GLMMs
library(visreg) #for plotting models
library(GGally) #for pairwise plotting
library(ggplot2) #for general plotting
library(readxl) #for reading in excel file
library(dplyr) #for dataframe management
library(plyr) #for dataframe management
library(car) #for ANOVA function
library(DHARMa) #for model validation
library(lmodel2)
library(lsmeans)
library(multcomp)
library(ggpubr)
library(arm)
library(BSDA)
library(tidyr)
Methods
(a) Fish Origin and Maintenance
We used laboratory-bred descendants of guppies from the Caura River in Trinidad. This laboratory population was founded by fish collected in 2012 from a high-predation population located at an undeveloped, rural site on the Caura River, Trinidad (UTM: 20 P 679527.7 m E, 1180376.4 m N based on WGS84 Datum; elevation 112m). Since 2012, the population has been maintained in large numbers, and we continue to see substantial variation in disease resistance, as is clear from our results. Fish were maintained and bred in 20 L tanks of mixed sexes under standard conditions, as follows. Tanks were on a recirculating system (Aquaneering) with approximately 20% daily water changes, a lighting schedule of 12L:12D, and an average temperature of 25±1°C. In these systems, wastewater from tanks through foam, sand, and ultraviolet filters before re-entering other tanks, preventing unintentional parasite transmission. Fish were fed flake and Artemia daily.
(b) Parasite origin and maintenance
Our parasite line was established by transferring one worm from a commercially obtained guppy to an uninfected host from a mixed laboratory-bred stock, a ‘culture fish’. This parasite line was maintained on groups of three to six culture fish per 1.8 L tank under standard conditions. Twice per week, all culture fish were anaesthetized (tricaine methanesulfonate ‘MS222’; 4g per L), and parasites were counted using a dissecting microscope. Uninfected fish were added to replace recovered or dead fish.
(c) Parental infection and breeding design
Parental fish were anaesthetized (using MS222 as above) and placed in close proximity to a dead, highly infected culture fish until at least two worms transmitted (mean±SEM=2.8±0.27), as observed under a dissecting microscope. Once infected, parental fish were isolated in 1.8 L tanks maintained under standard conditions for the duration of their infection. We counted worms every three days for 15 days to record individual infection trajectory over time. After 15 days, we treated fish with levamisole (0.002g per L) to clear infection. We verified that this treatment had cleared infection by screening the fish for parasites under anaesthetic, three times with a minimum of four days between screens. We recorded the length and weight of the parents before and after infection. Parasite naïve parents were exposed to the same conditions and manipulations, with the exception of parasites.
We set up 38 total breeding pairs of one male and one female using the recovered fish and parasite naïve fish in three different treatments: recovered mother, naïve father (13 pairs); recovered father, naïve mother (11 pairs); and both parents naïve (14 pairs). We did not set up pairs in which both parents were recovered because previous experience suggested this treatment may not have yielded sufficient offspring, and our primary objective was to examine the independent effects of maternal and paternal infection. These breeding pairs were housed in 1.8 L tanks under standard conditions, including fry mesh structures and gravel, to reduce possible fry cannibalism by the parents. Between May and December 2021, these breeding tanks were checked daily for the presence of offspring, which were moved to a separate 1.8 L tank. Each brood produced from a pair was moved to a separate 1.8 L tank.
(d) Offspring infection
Once they reached adulthood, offspring were separated into individual 1.8 L tanks. A total of 108 offspring were collected: 32 from recovered fathers and naïve mothers; 38 from recovered mothers and naïve fathers; and 38 from naïve parents. We used two batches of fish, one in August 2021 (52 offspring from 33 breeding pairs, with an average age at infection of approximately 4.4 months) and one in May 2022 (56 offspring from 32 breeding pairs, with an average age at infection of approximately 7.8 months). According to the needs of other concurrent experiments, 94 fish were infected manually, as described under parental infection, and 14 were infected via “co-housing”: they were housed in 1.8 L tanks with an infected conspecific and screened for infection every 24 hours. Once infected, these fish were isolated in a 1.8 L tank and treated the same as the manually infected offspring and parents from that point forward. As noted below, we included infection type as a fixed effect in the statistical models and found that it did not explain significant variation in our response variables. On all infected offspring, we counted the worms every three days for 12 days and then killed the host.