Data from: The effects of hybridization and parasite infection on predation survival and behaviour of endangered landlocked salmon – implications for genetic rescue
Data files
Jan 20, 2025 version files 1.89 MB
-
Antennadata_Daily_Summer.csv
279.82 KB
-
Antennadata_Daily_Winter.csv
692.99 KB
-
Antennadata_Hourly_Summer.csv
385.58 KB
-
Antennadata_Hourly_Winter.csv
420.52 KB
-
README.md
10 KB
-
Summer_experiment.csv
46.73 KB
-
Winter_experiment.csv
55.45 KB
Abstract
A prerequisite of genetic rescue in endangered and genetically depauperate populations is to pre-evaluate between possible pros and cons of hybridization for the life history and survival of the target population. We hybridized critically endangered Saimaa landlocked salmon (Salmo salar m. sebago) with one of its geographically closest relatives, anadromous Baltic salmon. In two experiments conducted in semi-natural streams during overwintering (at age 1.5 years) and in early summer (age 2+), we studied how hybridization and infection of eye parasites (Diplostomum pseudospathaceum) affected fish survival under predation from Northern pike (Esox lucius). Additionally, we recorded movements of the juvenile salmon to explore the effect of hybridization and infection on antipredatory behaviour (movement activity and habitat use). Among the uninfected groups, mortality of hybrid salmon was significantly lower (14.5%) compared to purebred landlocked salmon (37.2%), supporting a positive effect of hybridization under predation risk. This benefit, however, was removed by the parasite infection, which impaired vision and increased the susceptibility of fish to predation. The negative effects of infection were particularly pronounced in the anadromous salmon due to its lower infection resistance compared to the landlocked salmon. Hybridization per se did not affect activity levels of salmon, but overwintering activity correlated positively with eye cataract coverage, and summer activity was highest in anadromous salmon. These results demonstrate that controlled hybridization of a small animal population with a genetically more diverse population could entail both positive and negative implications, at least in the first crossbred generation.
README: The effects of hybridization and parasite infection on predation survival and behaviour of endangered landlocked salmon – implications for genetic rescue
https://doi.org/10.5061/dryad.qz612jmqk
In October 2018, we hybridized the critically endangered landlocked salmon of the Vuoksi watershed, Finland (Salmo salar m. sebago Girard, 1853) with a geographically close Baltic anadromous salmon (Salmo salar L. 1758) resulting in four crossing groups: purebred landlocked salmon (LL×LL), landlocked salmon females crossed with anadromous salmon males (LL×BA, hybrid), anadromous salmon females crossed with landlocked salmon males (BA×LL, hybrid), and purebred anadromous salmon (BA×BA).
Thereafter, we conducted two experiments in semi-natural stream tanks (circular channels, Fig. 1) to examine growth, predation-induced mortality and behavioral traits among the crossing groups. These experiments were conducted at two life stages: during overwintering (winter experiment, at age 1.5-2 years) and in early summer (summer experiment, at age 2+). In both experiments (winter and summer), half of the salmon were infected with the Diplostomum pseudospathaceum eye fluke on 20.2.2020 to explore how parasitism in a central sensory organ influenced the crossing groups.
Fish were measured before the start of the winter experiment on 26.10.-27.10.2020 and before the start of the summer experiment on 25.-26.5.2021. Half of the fish were placed in tanks that contained Northern pike (Esox lucius L. 1756) as piscine predators on the pool-side of the tanks, while the other half of the fish were placed in predator-free control tanks. Passive integrated telemetry (PIT) antennae were applied during the experiments to record the behaviour of individuals (i.e. activity and habitat use), both in the presence and absence of predators. The winter experiment was terminated after 154 days on 6./7.4.2021 and surviving fish were recorded and measured again. The summer experiment was terminated after 28 days on 28.6.2021. In the summer experiment survivors were recorded but not measured.
Description of the data and file structure
The data consists of the following three datasets (separated into summer and winter experiments):
- "Winter_experiment.csv" & "Summer_experiment.csv" consisting of the following columns:
PIT = Individual identification code of the salmon (Numeric code of Passive Integrated Transponder-microchip).
Strain = Crossing group, LLxLL, LLxBA, BAxLL or BAxBA (see above for details).
Infection = infection status (1 = Infected with D. pseudospathaceum, 0 = uninfected control).
Predation = The predation treatment (1 = in tank with piscine predators in pool-section, 0 = in predation-free control tank).
Tank = Number of circular tank, as depicted in Fig. 1 (there were eight tanks in total, four with pike, four without).
Length_start = Length in millimeters at the start of the experiment (For the winter experiment also Length*_*end is given).
Weight_start = Weight in grams at the start of the experiment (For the winter experiment also Weight*_*end is given).
Condition_start = Observed weight divided by expected weight (For the winter experiment also Condition*_*end is given).
cataract_l = Parasite-induced cataract coverage of the left eye (on a scale 0-10, where 0 = 0% coverage, 10 = 100% coverage).
cataract_r = Parasite-induced cataract coverage of the right eye (on a scale 0-10, where 0 = 0% coverage, 10 = 100% coverage).
DegreeofBlindness = cataract_l + cataract_r (used for calculating mean cataract coverage in both eye lenses).
FoundAlive = Result of the experiment, survival of the individual (0 = did not survive, 1 = was found alive, i.e. survived).
CauseofDeathPredation = Cause of death, as inferred from the PIT-telemetry data (0 = survived or died of unknown cause, 1 = eaten by pike).
Found_in = Where was the individual detected after the experiment (stream/pool-section of the tanks, or NA for the lost individuals).
POOLTIME = Time spent on the pool-section of the tanks (in minutes).
STREAMTIME = Time spent on the stream-section of the tanks (in minutes).
changes = Number of transitions between stream-and pool-section.
time.first.entry.to.pool = Time (in minutes) before the first entry from the stream-side (place of release) to the pool-section.
LastSignalIfDead = Last signal recorded (in minutes, can be used for inferring time of death).
halfrounds = Number of halfrounds completed, i.e. movements between two antenna pairs (1 halfround = ca. 10 meters).
NOTES = Notes on the individuals or the PIT-data.
Trial = experiment (winter or summer).
* "Antennadata_Daily*_*Winter.csv" & "Antennadata_Daily*_*Summer.csv"
PIT = Individual identification code the salmon.
Strain = Crossing group, LLxLL, LLxBA, BAxLL or BAxBA.
Infection = infection status (1 = Infected with D. pseudospathaceum, 0 = uninfected control).
Predation = The predation treatment (1 = in tank with piscine predators in pool-section, 0 = in predation-free control tank).
Tank = Number of circular tank, as depicted in Fig. 1(there were eight tanks in total, four with pike, four without).
DegreeofBlindness = cataract_l + cataract_r (used for calculating mean cataract coverage in both eye lenses).
FoundAlive = Result of the experiment, survival of the individual (0 = did not survive, 1 = was found alive, i.e. survived).
NOTES = Notes on the individuals or the PIT-data.
Time2 = The time (in days) from the start of the experiment, whence the individual was recorded (in winter 1-154 and in summer 1-28).
Total_day = The number of signals recorded per day (max. 1 signal/minute).
* "Antennadata_Hourly*_*Winter.csv" & "Antennadata_Hourly*_*Summer.csv"
PIT = Individual identification code the salmon.
Strain = Crossing group, LLxLL, LLxBA, BAxLL or BAxBA.
Infection = infection status (1 = Infected with D. pseudospathaceum, 0 = uninfected control).
Predation = The predation treatment (1 = in tank with piscine predators in pool-section, 0 = in predation-free control tank).
Tank = Number of circular tank, as depicted in Fig. 1 (there were eight tanks in total, four with pike, four without).
DegreeofBlindness = cataract_l + cataract_r (used for calculating mean cataract coverage in both eye lenses).
FoundAlive = Result of the experiment, survival of the individual (0 = did not survive, 1 = was found alive, i.e. survived).
NOTES = Notes on the individuals or the PIT-data.
Time3 = The circadian time (hour) when the individual was recorded (0-23, where 0 = 00:00 to 00:59, 1 = 1:00 to 01:59 etc.).
Total_hour = the number of signals recorded per hour (max. 1 signal/minute).
Code/Software
R-code used for analysis of this data will be attached. The analysis consists of the following five codes:
- "1_Preliminary_Analyses*_*Winter" & "2_Preliminary_Analyses*_*Summer"
Analyses on the size (lengths and weights) and D. pseudospathaceum-induced cataract coverage of the eye lens (on a scale 0-10, where 0 = 0% coverage, 10 = 100% coverage) of the fish involved, as measured before the start of the experiments. Calculation of the condition indexes. Analysis of Variance with Tukey-adjusted post-hoc comparisons is used for comparing sizes, cataract coverage and condition indexes among crossing groups and infection treatments (1 = Infected with D. pseudospathaceum, 0 = uninfected control). For the winter experiment also sizes and condition indexes of the surviving individuals after the experiment are analyzed taking into account the predation treatment (1 = in tank with piscine predators in pool-section, 0 = in predation-free control tank). R-packages used: "dplyr" & "mgcv".
* "3_Mortality_Analyses"
Analyses of predation-induced mortality among the fish of the predation tanks (binomial cause of death, as inferred from the PIT-telemetry data: 0 = survived or died of unknown cause, 1 = eaten by pike). Data of the winter and summer experiments is combined for these analyses. The data were analysed at three stages using generalized linear mixed models with a binomial error distribution: 1. Predation-induced mortality among all fish of the predation tanks, accounting for the binomial infection status (1 = Infected with D. pseudospathaceum, 0 = uninfected control), 2. Predation-induced mortality among uninfected fish only (infection status 0 in the first model), and 3. Predation-induced mortality among infected fish only (infection status 1 in the first model). Model factors and variance structures were chosen based on Akaike information criteria (AIC-values) starting from the most complex models with all interactions and excluding interactions and factors one by one and paired post hoc comparisons were based on Šidák corrected values. R-packages used: "dplyr", "tidyr", "mgcv", "glmmTMB", "DHARMa", "emmeans","multcomp" & "ggplot2".
* "4_Movement_Analyses*_*Winter_excluding_J505" & "5_Movement_Analyses*_*Summer"
Analysis of the behaviour (activity and habitat use) of the surviving individuals, as inferred from the PIT-telemetry-data. In the winter experiment tank J505 is excluded from the analysis because the gate preventing movement between stream- and pool-sections had opened and some of the fish went to the pool-section before the PIT-antennae were started. These analyses consist of three parts: 1. A Principal Components Analysis (PCA) with varimax-rotation is conducted to reduce the six behavioural variables into two principal components, 2. Analyzing PC1 (reflecting overall activity) using plots and linear mixed models, and 3. Analyzing PC2 (rweflecting habitat use and circadian rhythm) using plots and linear mixed models. R-packages used: "dplyr", "vegan", "ggfortify", "nlme", "lmtest", "mgcv", "effectsize", "multcomp", "emmeans" & "sjPlot".
Methods
In October 2018, we hybridized the critically endangered landlocked salmon of the Vuoksi watershed, Finland (Salmo salar m. sebago Girard, 1853) with a geographically close Baltic anadromous salmon (Salmo salar L. 1758, population spawning in river Kymijoki, Finland, but originally from river Neva, Russia). Parental fish consisted of 15 males and 15 females from both populations, with eggs from each female being fertilized with one male from its own population and the other part with one male from the other population. Correspondingly, equal portions of milt from each male were used to fertilize eggs of one female from its own population and of one female from the other population (n = 60 full-sibling families, 15 per crossing group). This resulted in four crossing groups: purebred landlocked salmon (LL×LL), landlocked salmon females crossed with anadromous salmon males (LL×BA, hybrid), anadromous salmon females crossed with landlocked salmon males (BA×LL, hybrid), and purebred anadromous salmon (BA×BA).
Thereafter, we conducted two experiments in semi-natural circular stream tanks (Fig. 1) to examine growth, predation-induced mortality and behavioural traits among the crossing groups. These experiments were conducted at two life stages: during overwintering (winter experiment, at age 1.5-2 years) and in early summer (summer experiment, at age 2+). In both experiments (winter and summer), half of the salmon were infected with the D. pseudospathaceum eye fluke on 20.2.2020 to explore how parasitism in a central sensory organ influenced the crossing groups. Fish were measured before the start of the winter experiment on 26.10.-27.10.2020 and before the start of the summer experiment on 25.-26.5.2021.
After measuring the fish were transferred into the shallow stream-section of the experimental tanks (Fig. 1) on the 27.10.2020 (winter experiment) and 27.5.2021 (summer experiment). Into the pool-sections of tanks J506, J507, J513 and J516 we transferred in the winter experiment two Northern pike (Esox Lucius Linnaeus 1758, common piscine predators of juvenile salmonids) and in the summer experiment three pikes per tank. Tanks J505, J508, J514 and J515 were control tanks without predators. The winter experiment was terminated after 154 days on 6./7.4.2021 and surviving fish were recorded and measured again. The summer experiment was terminated after 28 days on 28.6.2021 when mortality was estimated to be similar than in the winter experiment. In the summer experiment survivors were recorded but not measured.
Passive integrated telemetry (PIT) antennae were applied during the experiments to record the behaviour of individuals (i.e. activity and habitat use), both in the presence and absence of predators. PIT-antennae were started and gates opened 3.11.2020 (winter experiment) and 31.5.2021 (summer experiment), enabling the salmon to move to the potential predators on the pool-sides. In the winter experiments tanks J513 and J505 the gates had opened before the PIT antennae were started (PIT-data potentially unusable).
The study was conducted at the Kainuu Fisheries Research Station in Paltamo, Finland (64ᵒ24’N 27ᵒ31’E) maintained by the Natural Resources Institute Finland (Luke).