Computational protocol: The signature of fine scale local adaptation in Atlantic salmon revealed from common garden experiments in nature

Similar protocols

Protocol publication

[…] Variation in the length and mass of individual offspring was also examined in separate analyses for each of the five life/sampling stages. Linear mixed effects models (LMMs) were used assuming normally distributed errors and including dam and sire as random effects. Fixed effects of group: mean eyed-egg diameter, dam LF, date of egg-planting and, where appropriate, date of capture and its square (to capture nonlinear growth patterns) were included as candidate explanatory variables in all models. Date of egg-planting was a three level factor: ‘early’ = 2nd March 2009 (65% of families), ‘mid’ = 9th March 2009 (27% of families), ‘late’ = 16th and 20th March 2009 (8% of families). Since egg diameter and egg volume were highly correlated (r = 0.87) only egg diameter was included in the LMMs to avoid problems associated with multicollinearity (the correlation between dam LF and mean eyed-egg diameter was 0.32). Nonsignificant fixed effects (as determined by dropping terms one at a time and using a likelihood ratio test (LRTs) to compare nested models fit by maximum likelihood) were removed in turn to identify the minimum adequate model (Zuur et al. ). Results were in all cases robust to different random effect structures (dropping dam or sire or both, structuring the residual variance by group or leaving unstructured). LMMs were fit in R version 3.0.2 (R Core Development Team 2008) using the lme function from the nlme package (Pinheiro et al. ), and all continuous covariates were first z-standardised. […]

Pipeline specifications

Software tools lme4, nlme
Application Mathematical modeling
Organisms Salmo salar