7 Eisen und Zink in Salat

Letzte Änderung am 13. January 2024 um 17:46:55

Genutzte R Pakete

pacman::p_load(tidyverse, readxl, parameters,
               effectsize, magrittr, multcomp,
               multcompView, rcompanion, rstatix,
               emmeans, see, performance, fs,
               janitor, broom, scales, knitr,
               conflicted)
## resolve some conflicts with same function naming
conflicts_prefer(dplyr::select)
conflicts_prefer(dplyr::filter)
conflicts_prefer(effectsize::eta_squared)
conflicts_prefer(magrittr::set_names)

7.1 Daten einlesen

data_tbl <- read_csv2("data/fe_zn_salad.csv") %>% 
  clean_names() %>% 
  fill(eisendungeform) %>% 
  mutate(eisendungeform = str_replace_all(eisendungeform, "\u0096 7 ", ""),
         versuchsgruppe = str_c(eisendungeform, "_", eisendungung)) %>% 
  pivot_longer(cols = zn_blanchiert:fe_ungewaschen,
               names_to = "outcome",
               values_to = "rsp")

7.2 Daten auswerten

nested_tbl <- data_tbl %>% 
  select(versuchsgruppe, outcome, rsp) %>% 
  group_by(outcome) %>% 
  nest() 

aov_res <- nested_tbl %>% 
  mutate(model = map(data, ~lm(rsp ~ versuchsgruppe, data = .x))) %>% 
  mutate(anova = map(model, anova)) %>% 
  mutate(parameter = map(anova, model_parameters)) %>% 
  select(outcome, parameter) %>% 
  unnest(parameter) %>% 
  clean_names() %>% 
  mutate(p = pvalue(p)) %>% 
  filter(parameter != "Residuals") %>% 
  select(outcome, parameter, p)

aov_res %>% 
  kable(align = "c", "pipe")

outcome	parameter	p
zn_blanchiert	versuchsgruppe	0.538
zn_ungewaschen	versuchsgruppe	0.634
fe_blanchiert	versuchsgruppe	<0.001
fe_ungewaschen	versuchsgruppe	<0.001

emm_tbl <- nested_tbl %>% 
  mutate(model = map(data, ~lm(rsp ~ versuchsgruppe, data = .x))) %>% 
  mutate(emm = map(model, emmeans, ~versuchsgruppe)) %>% 
  mutate(contrast = map(emm, contrast, method = "pairwise", 
                        adjust = "none")) %>% 
  mutate(cld = map(emm, cld, Letters = letters, adjust = "none"))
  
emm_tbl %>% 
  mutate(contrast = map(contrast, as_tibble)) %>% 
  unnest(contrast) %>% 
  select(outcome, contrast, p.value) %>% 
  mutate(p.value = pvalue(p.value)) %>% 
  print(n = 10)

# A tibble: 144 × 3
# Groups:   outcome [4]
   outcome       contrast                        p.value
   <chr>         <chr>                           <chr>  
 1 zn_blanchiert (Fe-EDDHA_0.2) - (Fe-EDDHA_0.3) 0.836  
 2 zn_blanchiert (Fe-EDDHA_0.2) - (Fe-EDDHA_0.4) 0.118  
 3 zn_blanchiert (Fe-EDDHA_0.2) - (Fe-EDDHA_0.6) 0.357  
 4 zn_blanchiert (Fe-EDDHA_0.2) - FeSO4 H2O_0.2  0.682  
 5 zn_blanchiert (Fe-EDDHA_0.2) - FeSO4 H2O_0.3  0.573  
 6 zn_blanchiert (Fe-EDDHA_0.2) - FeSO4 H2O_0.4  0.878  
 7 zn_blanchiert (Fe-EDDHA_0.2) - FeSO4 H2O_0.6  0.492  
 8 zn_blanchiert (Fe-EDDHA_0.2) - Kontrolle_0    0.168  
 9 zn_blanchiert (Fe-EDDHA_0.3) - (Fe-EDDHA_0.4) 0.171  
10 zn_blanchiert (Fe-EDDHA_0.3) - (Fe-EDDHA_0.6) 0.474  
# ℹ 134 more rows

emm_tbl %>% 
  mutate(cld = map(cld, arrange, versuchsgruppe)) %>% 
  unnest(cld) %>% 
  select(outcome, versuchsgruppe, .group) %>% 
  print(n = Inf)

# A tibble: 36 × 3
# Groups:   outcome [4]
   outcome        versuchsgruppe .group  
   <chr>          <fct>          <chr>   
 1 zn_blanchiert  Fe-EDDHA_0.2   " a"    
 2 zn_blanchiert  Fe-EDDHA_0.3   " a"    
 3 zn_blanchiert  Fe-EDDHA_0.4   " a"    
 4 zn_blanchiert  Fe-EDDHA_0.6   " a"    
 5 zn_blanchiert  FeSO4 H2O_0.2  " a"    
 6 zn_blanchiert  FeSO4 H2O_0.3  " a"    
 7 zn_blanchiert  FeSO4 H2O_0.4  " a"    
 8 zn_blanchiert  FeSO4 H2O_0.6  " a"    
 9 zn_blanchiert  Kontrolle_0    " a"    
10 zn_ungewaschen Fe-EDDHA_0.2   " a"    
11 zn_ungewaschen Fe-EDDHA_0.3   " a"    
12 zn_ungewaschen Fe-EDDHA_0.4   " a"    
13 zn_ungewaschen Fe-EDDHA_0.6   " a"    
14 zn_ungewaschen FeSO4 H2O_0.2  " a"    
15 zn_ungewaschen FeSO4 H2O_0.3  " a"    
16 zn_ungewaschen FeSO4 H2O_0.4  " a"    
17 zn_ungewaschen FeSO4 H2O_0.6  " a"    
18 zn_ungewaschen Kontrolle_0    " a"    
19 fe_blanchiert  Fe-EDDHA_0.2   "  bc  "
20 fe_blanchiert  Fe-EDDHA_0.3   "  bc  "
21 fe_blanchiert  Fe-EDDHA_0.4   "   cd "
22 fe_blanchiert  Fe-EDDHA_0.6   "    d "
23 fe_blanchiert  FeSO4 H2O_0.2  " ab   "
24 fe_blanchiert  FeSO4 H2O_0.3  "    d "
25 fe_blanchiert  FeSO4 H2O_0.4  "    d "
26 fe_blanchiert  FeSO4 H2O_0.6  "     e"
27 fe_blanchiert  Kontrolle_0    " a    "
28 fe_ungewaschen Fe-EDDHA_0.2   "  b   "
29 fe_ungewaschen Fe-EDDHA_0.3   "  bc  "
30 fe_ungewaschen Fe-EDDHA_0.4   "   cd "
31 fe_ungewaschen Fe-EDDHA_0.6   "    d "
32 fe_ungewaschen FeSO4 H2O_0.2  "  b   "
33 fe_ungewaschen FeSO4 H2O_0.3  "  bc  "
34 fe_ungewaschen FeSO4 H2O_0.4  "  bc  "
35 fe_ungewaschen FeSO4 H2O_0.6  "     e"
36 fe_ungewaschen Kontrolle_0    " a    "

gg_func <- function(data){
  data %>% 
    mutate(.group = str_replace_all(.group, "\\s", "")) %>% 
    ggplot(aes(versuchsgruppe, emmean)) +
    theme_minimal() +
    geom_bar(stat = "identity")  +
    geom_errorbar(aes(ymin = emmean - SE, ymax = emmean + SE), width = 0.2) +
    geom_text(aes(label = .group, y = (emmean + SE) + (0.05 * (emmean + SE)))) +
    theme(axis.text.x = element_text(angle = -45, hjust = 0))
}

Plot math expression

emm_tbl %>% 
  mutate(gg = map(cld, gg_func)) %>% 
  pluck("gg", 1) +
  labs(x = expression(Eisendüngeform~und~-höhe~"[kg ha]"^-1),
       y = expression(Zn-Gehalt~"["~mg%.%(100~g~FM)^-1~"]"))

emm_tbl %>% 
  mutate(gg = map(cld, gg_func)) %>% 
  pluck("gg", 2) +
  labs(x = expression(Eisendüngeform~und~-höhe~"[kg ha]"^-1),
       y = expression(Zn-Gehalt~"["~mg%.%(kg~TM)^-1~"]"))

emm_tbl %>% 
  mutate(gg = map(cld, gg_func)) %>% 
  pluck("gg", 3) +
  labs(x = expression(Eisendüngeform~und~-höhe~"[kg ha]"^-1),
       y = expression(Fe-Gehalt~"["~mg%.%(100~g~FM)^-1~"]"))

emm_tbl %>% 
  mutate(gg = map(cld, gg_func)) %>% 
  pluck("gg", 4) +
  labs(x = expression(Eisendüngeform~und~-höhe~"[kg ha]"^-1),
       y = expression(Fe-Gehalt~"["~mg%.%(kg~TM)^-1~"]"))