A0327
Title: Applying regression methods to model survival data for gammarids (Amphipoda)
Authors: Svitlana Shvydka - Slovak University of Technology in Bratislava (Slovakia) [presenting]
Volodimir Sarabeev - Institute of Parasitology (Slovakia)
Maria Zdimalova - Slovak University of Technology in Bratislava (Slovakia)
Mykola Ovcharenko - Institute of Biology and Science of Earth (Poland)
Abstract: The study considers the survival time of gammarids (Amphipoda) using linear and nonlinear regression models with Gaussian distribution. The effect of individual parasite species richness (IPSR) on survival data under the influence of the host body length was analyzed. Three models were tested: GLM, Survival_time~Host_species+IPSR:Body_length; GAM, Survival_time~Host_species+s(IPSR)+s(Body_length)+ti(IPSR,Body_length); ANCOVA, Survival_time~IPSR+Body_length. For GLM and GAM models, the coefficients of the interaction terms were not significant. The marginal effects for both models revealed that the IPSR affected the survival time of gammarids measuring 7.7 to 15 mm in body length, but this influence was not observed in larger hosts. The smoother of Body_length in GAM showed a non-linear effect. Gammarids around 11 mm have a higher probability of having a number of parasite species than smaller and larger hosts. The ANCOVA showed a significant difference in the effect of parasites on survival time between groups of hosts with body length of 7.7-15 and 15-21.8 mm. We evaluated the model fit by checking the level of deviance explained, analyzing the AIC and inspecting the residuals. S.S., V.S. are funded by the EU NextGenerationEU through the Recovery and Resilience Plan for Slovakia under project Nos.09I03-03-V01-00029, 09I03-03-V01-00017, respectively. The work of S.S. is partially supported by the grant VEGA 1/0036/23.
