setwd("~/Library/CloudStorage/GoogleDrive-icarounam@gmail.com/Mi unidad/Agrosavia/Colaboraciones/René/1er_Cd/data")
datos<-read.table("percentage2.csv", header=T, sep=',')
datos$curva <- factor(datos$curva, levels = c("1", "2", "3"),
labels = c("T3", "T1", "T2"))
datos$gen<-as.factor(datos$gen)
datos$curva<-as.factor(datos$curva)
datos$id<-as.factor(datos$id)
datos$muestra<-as.factor(datos$muestra)
datos$diam2<-as.factor(datos$diam2)
library(ggplot2)
library(Rmisc)
## Loading required package: lattice
## Loading required package: plyr
library(dplyr)
## Warning: package 'dplyr' was built under R version 4.1.2
##
## Attaching package: 'dplyr'
## The following objects are masked from 'package:plyr':
##
## arrange, count, desc, failwith, id, mutate, rename, summarise,
## summarize
## The following objects are masked from 'package:stats':
##
## filter, lag
## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, union
library(tidyverse)
## ── Attaching packages ─────────────────────────────────────── tidyverse 1.3.1 ──
## ✔ tibble 3.2.1 ✔ purrr 1.0.1
## ✔ tidyr 1.3.0 ✔ stringr 1.5.0
## ✔ readr 2.1.1 ✔ forcats 1.0.0
## Warning: package 'tibble' was built under R version 4.1.2
## Warning: package 'tidyr' was built under R version 4.1.2
## Warning: package 'purrr' was built under R version 4.1.2
## Warning: package 'stringr' was built under R version 4.1.2
## Warning: package 'forcats' was built under R version 4.1.2
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::arrange() masks plyr::arrange()
## ✖ purrr::compact() masks plyr::compact()
## ✖ dplyr::count() masks plyr::count()
## ✖ dplyr::desc() masks plyr::desc()
## ✖ dplyr::failwith() masks plyr::failwith()
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::id() masks plyr::id()
## ✖ dplyr::lag() masks stats::lag()
## ✖ dplyr::mutate() masks plyr::mutate()
## ✖ dplyr::rename() masks plyr::rename()
## ✖ dplyr::summarise() masks plyr::summarise()
## ✖ dplyr::summarize() masks plyr::summarize()
library(ggpubr)
##
## Attaching package: 'ggpubr'
## The following object is masked from 'package:plyr':
##
## mutate
library(rstatix)
##
## Attaching package: 'rstatix'
## The following objects are masked from 'package:plyr':
##
## desc, mutate
## The following object is masked from 'package:stats':
##
## filter
library(emmeans)
## ITF
##Summary statistics
summ<-datos %>%
group_by(curva, gen) %>%
get_summary_stats(tf, type = "mean_sd")
summ %>% as_tibble() %>% print(n=Inf)
## # A tibble: 9 × 6
## curva gen variable n mean sd
## <fct> <fct> <chr> <dbl> <dbl> <dbl>
## 1 T3 CCN51 tf 9 1.23 0.441
## 2 T3 ICS95 tf 9 1.14 0.331
## 3 T3 TCS01 tf 9 0.871 0.262
## 4 T1 CCN51 tf 9 1.53 0.396
## 5 T1 ICS95 tf 9 0.966 0.142
## 6 T1 TCS01 tf 9 0.962 0.192
## 7 T2 CCN51 tf 9 1.26 0.402
## 8 T2 ICS95 tf 9 0.81 0.132
## 9 T2 TCS01 tf 9 0.916 0.323
##Visualization
bxp <- ggboxplot(
datos, x = "curva", y = "tf",
palette = "jco",
color = "gen", xlab = "Treatment", ylab = "ITF", legend.title = "Genotype"
)
bxp
##Check assumptions
##Outliers
datos %>%
group_by(curva, gen) %>%
identify_outliers(tf)
## # A tibble: 1 × 21
## curva gen ids trat muestra id diam2 cd.testa cd.grano pdcd.grano
## <fct> <fct> <int> <chr> <fct> <fct> <fct> <dbl> <dbl> <dbl>
## 1 T2 ICS95 88 T2 14 22 6 11.3 10.2 -0.183
## # ℹ 11 more variables: picd.testa <dbl>, pdph.grano <dbl>, piph.testa <dbl>,
## # ph.testa <dbl>, acidez.testa <dbl>, ph.grano <dbl>, acidez.grano <dbl>,
## # tf <dbl>, pi.tf <dbl>, is.outlier <lgl>, is.extreme <lgl>
##Normality assumption
##Compute Shapiro-Wilk test for each combinations of factor levels:
norm<-datos %>%
group_by(curva, gen) %>%
shapiro_test(tf)
norm %>% as_tibble() %>% print(n=Inf)
## # A tibble: 9 × 5
## curva gen variable statistic p
## <fct> <fct> <chr> <dbl> <dbl>
## 1 T3 CCN51 tf 0.916 0.357
## 2 T3 ICS95 tf 0.896 0.228
## 3 T3 TCS01 tf 0.973 0.917
## 4 T1 CCN51 tf 0.937 0.547
## 5 T1 ICS95 tf 0.887 0.187
## 6 T1 TCS01 tf 0.966 0.863
## 7 T2 CCN51 tf 0.806 0.0238
## 8 T2 ICS95 tf 0.869 0.121
## 9 T2 TCS01 tf 0.920 0.394
##Create QQ plot for each cell of design:
ggqqplot(datos, "tf", ggtheme = theme_bw()) +
facet_grid(curva~ gen, labeller = "label_both")
## Warning: The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## Warning: The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
##Homogneity of variance assumption
##Compute the Levene’s test at each level of the within-subjects factor, here time variable:
lev<-datos %>%
group_by(curva) %>%
levene_test(tf ~ gen)
lev %>% as_tibble() %>% print(n=Inf)
## # A tibble: 3 × 5
## curva df1 df2 statistic p
## <fct> <int> <int> <dbl> <dbl>
## 1 T3 2 24 1.07 0.359
## 2 T1 2 24 3.81 0.0366
## 3 T2 2 24 2.09 0.146
##Computation
res.aov <- anova_test(
data = datos, dv = tf, wid = id,
within = diam2, between = c(curva, gen)
)
res.aov
## ANOVA Table (type II tests)
##
## $ANOVA
## Effect DFn DFd F p p<.05 ges
## 1 curva 2 18 6.083 1.00e-02 * 0.241
## 2 gen 2 18 53.431 2.69e-08 * 0.736
## 3 diam2 2 36 156.491 1.75e-18 * 0.822
## 4 curva:gen 4 18 6.277 2.00e-03 * 0.396
## 5 curva:diam2 4 36 3.015 3.00e-02 * 0.151
## 6 gen:diam2 4 36 11.210 5.21e-06 * 0.397
## 7 curva:gen:diam2 8 36 2.873 1.40e-02 * 0.253
##
## $`Mauchly's Test for Sphericity`
## Effect W p p<.05
## 1 diam2 0.983 0.867
## 2 curva:diam2 0.983 0.867
## 3 gen:diam2 0.983 0.867
## 4 curva:gen:diam2 0.983 0.867
##
## $`Sphericity Corrections`
## Effect GGe DF[GG] p[GG] p[GG]<.05 HFe DF[HF]
## 1 diam2 0.984 1.97, 35.41 3.24e-18 * 1.103 2.21, 39.72
## 2 curva:diam2 0.984 3.93, 35.41 3.10e-02 * 1.103 4.41, 39.72
## 3 gen:diam2 0.984 3.93, 35.41 6.11e-06 * 1.103 4.41, 39.72
## 4 curva:gen:diam2 0.984 7.87, 35.41 1.50e-02 * 1.103 8.83, 39.72
## p[HF] p[HF]<.05
## 1 1.75e-18 *
## 2 3.00e-02 *
## 3 5.21e-06 *
## 4 1.40e-02 *
get_anova_table(res.aov)
## ANOVA Table (type II tests)
##
## Effect DFn DFd F p p<.05 ges
## 1 curva 2 18 6.083 1.00e-02 * 0.241
## 2 gen 2 18 53.431 2.69e-08 * 0.736
## 3 diam2 2 36 156.491 1.75e-18 * 0.822
## 4 curva:gen 4 18 6.277 2.00e-03 * 0.396
## 5 curva:diam2 4 36 3.015 3.00e-02 * 0.151
## 6 gen:diam2 4 36 11.210 5.21e-06 * 0.397
## 7 curva:gen:diam2 8 36 2.873 1.40e-02 * 0.253
#Table by error
res.aov.error <- aov(tf ~ diam2*curva*gen + Error(id/diam2), datos)
res.aov.error
##
## Call:
## aov(formula = tf ~ diam2 * curva * gen + Error(id/diam2), data = datos)
##
## Grand Mean: 1.076272
##
## Stratum 1: id
##
## Terms:
## curva gen curva:gen Residuals
## Sum of Squares 0.3267770 2.8702724 0.6743588 0.4834690
## Deg. of Freedom 2 2 4 18
##
## Residual standard error: 0.1638883
## 16 out of 24 effects not estimable
## Estimated effects may be unbalanced
##
## Stratum 2: id:diam2
##
## Terms:
## diam2 diam2:curva diam2:gen diam2:curva:gen Residuals
## Sum of Squares 4.729690 0.182254 0.677613 0.347289 0.544020
## Deg. of Freedom 2 4 4 8 36
##
## Residual standard error: 0.1229295
## Estimated effects may be unbalanced
summary(res.aov.error)
##
## Error: id
## Df Sum Sq Mean Sq F value Pr(>F)
## curva 2 0.3268 0.1634 6.083 0.00959 **
## gen 2 2.8703 1.4351 53.431 2.69e-08 ***
## curva:gen 4 0.6744 0.1686 6.277 0.00241 **
## Residuals 18 0.4835 0.0269
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Error: id:diam2
## Df Sum Sq Mean Sq F value Pr(>F)
## diam2 2 4.730 2.3648 156.491 < 2e-16 ***
## diam2:curva 4 0.182 0.0456 3.015 0.0304 *
## diam2:gen 4 0.678 0.1694 11.210 5.21e-06 ***
## diam2:curva:gen 8 0.347 0.0434 2.873 0.0140 *
## Residuals 36 0.544 0.0151
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## Emmeans
emmip(res.aov.error, gen ~ diam2 | curva)
## Note: re-fitting model with sum-to-zero contrasts
emm_curva <- emmeans(res.aov.error, pairwise ~ curva)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_curva
## $emmeans
## curva emmean SE df lower.CL upper.CL
## T3 1.081 0.0315 18 1.02 1.15
## T1 1.151 0.0315 18 1.09 1.22
## T2 0.996 0.0315 18 0.93 1.06
##
## Results are averaged over the levels of: diam2, gen
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## T3 - T1 -0.0701 0.0446 18 -1.573 0.2826
## T3 - T2 0.0852 0.0446 18 1.910 0.1646
## T1 - T2 0.1553 0.0446 18 3.483 0.0071
##
## Results are averaged over the levels of: diam2, gen
## P value adjustment: tukey method for comparing a family of 3 estimates
emm_gen_curva <- emmeans(res.aov.error, pairwise ~ gen | curva)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_gen_curva
## $emmeans
## curva = T3:
## gen emmean SE df lower.CL upper.CL
## CCN51 1.233 0.0546 18 1.118 1.348
## ICS95 1.140 0.0546 18 1.026 1.255
## TCS01 0.871 0.0546 18 0.756 0.985
##
## curva = T1:
## gen emmean SE df lower.CL upper.CL
## CCN51 1.527 0.0546 18 1.412 1.642
## ICS95 0.966 0.0546 18 0.851 1.080
## TCS01 0.962 0.0546 18 0.847 1.077
##
## curva = T2:
## gen emmean SE df lower.CL upper.CL
## CCN51 1.262 0.0546 18 1.147 1.377
## ICS95 0.810 0.0546 18 0.695 0.925
## TCS01 0.916 0.0546 18 0.802 1.031
##
## Results are averaged over the levels of: diam2
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## curva = T3:
## contrast estimate SE df t.ratio p.value
## CCN51 - ICS95 0.09258 0.0773 18 1.198 0.4692
## CCN51 - TCS01 0.36237 0.0773 18 4.690 0.0005
## ICS95 - TCS01 0.26978 0.0773 18 3.492 0.0070
##
## curva = T1:
## contrast estimate SE df t.ratio p.value
## CCN51 - ICS95 0.56119 0.0773 18 7.264 <.0001
## CCN51 - TCS01 0.56473 0.0773 18 7.310 <.0001
## ICS95 - TCS01 0.00354 0.0773 18 0.046 0.9988
##
## curva = T2:
## contrast estimate SE df t.ratio p.value
## CCN51 - ICS95 0.45195 0.0773 18 5.850 <.0001
## CCN51 - TCS01 0.34560 0.0773 18 4.473 0.0008
## ICS95 - TCS01 -0.10635 0.0773 18 -1.377 0.3735
##
## Results are averaged over the levels of: diam2
## P value adjustment: tukey method for comparing a family of 3 estimates
emm_gen_diam2 <- emmeans(res.aov.error, pairwise ~ curva*gen)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_gen_diam2
## $emmeans
## curva gen emmean SE df lower.CL upper.CL
## T3 CCN51 1.233 0.0546 18 1.118 1.348
## T1 CCN51 1.527 0.0546 18 1.412 1.642
## T2 CCN51 1.262 0.0546 18 1.147 1.377
## T3 ICS95 1.140 0.0546 18 1.026 1.255
## T1 ICS95 0.966 0.0546 18 0.851 1.080
## T2 ICS95 0.810 0.0546 18 0.695 0.925
## T3 TCS01 0.871 0.0546 18 0.756 0.985
## T1 TCS01 0.962 0.0546 18 0.847 1.077
## T2 TCS01 0.916 0.0546 18 0.802 1.031
##
## Results are averaged over the levels of: diam2
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## T3 CCN51 - T1 CCN51 -0.29380 0.0773 18 -3.803 0.0276
## T3 CCN51 - T2 CCN51 -0.02901 0.0773 18 -0.375 1.0000
## T3 CCN51 - T3 ICS95 0.09258 0.0773 18 1.198 0.9467
## T3 CCN51 - T1 ICS95 0.26739 0.0773 18 3.461 0.0544
## T3 CCN51 - T2 ICS95 0.42295 0.0773 18 5.474 0.0009
## T3 CCN51 - T3 TCS01 0.36237 0.0773 18 4.690 0.0045
## T3 CCN51 - T1 TCS01 0.27093 0.0773 18 3.507 0.0497
## T3 CCN51 - T2 TCS01 0.31659 0.0773 18 4.098 0.0152
## T1 CCN51 - T2 CCN51 0.26480 0.0773 18 3.427 0.0580
## T1 CCN51 - T3 ICS95 0.38639 0.0773 18 5.001 0.0023
## T1 CCN51 - T1 ICS95 0.56119 0.0773 18 7.264 <.0001
## T1 CCN51 - T2 ICS95 0.71675 0.0773 18 9.277 <.0001
## T1 CCN51 - T3 TCS01 0.65617 0.0773 18 8.493 <.0001
## T1 CCN51 - T1 TCS01 0.56473 0.0773 18 7.310 <.0001
## T1 CCN51 - T2 TCS01 0.61040 0.0773 18 7.901 <.0001
## T2 CCN51 - T3 ICS95 0.12159 0.0773 18 1.574 0.8064
## T2 CCN51 - T1 ICS95 0.29639 0.0773 18 3.836 0.0258
## T2 CCN51 - T2 ICS95 0.45195 0.0773 18 5.850 0.0004
## T2 CCN51 - T3 TCS01 0.39137 0.0773 18 5.066 0.0020
## T2 CCN51 - T1 TCS01 0.29994 0.0773 18 3.882 0.0235
## T2 CCN51 - T2 TCS01 0.34560 0.0773 18 4.473 0.0070
## T3 ICS95 - T1 ICS95 0.17480 0.0773 18 2.263 0.4105
## T3 ICS95 - T2 ICS95 0.33036 0.0773 18 4.276 0.0105
## T3 ICS95 - T3 TCS01 0.26978 0.0773 18 3.492 0.0512
## T3 ICS95 - T1 TCS01 0.17835 0.0773 18 2.308 0.3861
## T3 ICS95 - T2 TCS01 0.22401 0.0773 18 2.900 0.1536
## T1 ICS95 - T2 ICS95 0.15556 0.0773 18 2.014 0.5536
## T1 ICS95 - T3 TCS01 0.09498 0.0773 18 1.229 0.9390
## T1 ICS95 - T1 TCS01 0.00354 0.0773 18 0.046 1.0000
## T1 ICS95 - T2 TCS01 0.04921 0.0773 18 0.637 0.9990
## T2 ICS95 - T3 TCS01 -0.06058 0.0773 18 -0.784 0.9960
## T2 ICS95 - T1 TCS01 -0.15201 0.0773 18 -1.968 0.5812
## T2 ICS95 - T2 TCS01 -0.10635 0.0773 18 -1.377 0.8927
## T3 TCS01 - T1 TCS01 -0.09144 0.0773 18 -1.184 0.9502
## T3 TCS01 - T2 TCS01 -0.04577 0.0773 18 -0.592 0.9994
## T1 TCS01 - T2 TCS01 0.04566 0.0773 18 0.591 0.9994
##
## Results are averaged over the levels of: diam2
## P value adjustment: tukey method for comparing a family of 9 estimates
emm_gen_diam2_trend <- emmeans(res.aov.error, pairwise ~ gen)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_gen_diam2_trend
## $emmeans
## gen emmean SE df lower.CL upper.CL
## CCN51 1.341 0.0315 18 1.274 1.407
## ICS95 0.972 0.0315 18 0.906 1.038
## TCS01 0.916 0.0315 18 0.850 0.983
##
## Results are averaged over the levels of: diam2, curva
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## CCN51 - ICS95 0.3686 0.0446 18 8.263 <.0001
## CCN51 - TCS01 0.4242 0.0446 18 9.511 <.0001
## ICS95 - TCS01 0.0557 0.0446 18 1.248 0.4416
##
## Results are averaged over the levels of: diam2, curva
## P value adjustment: tukey method for comparing a family of 3 estimates
emm_diam2 <- emmeans(res.aov.error, pairwise ~ diam2)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_diam2
## $emmeans
## diam2 emmean SE df lower.CL upper.CL
## 2 0.753 0.0265 49.8 0.699 0.806
## 5 1.143 0.0265 49.8 1.089 1.196
## 6 1.333 0.0265 49.8 1.280 1.387
##
## Results are averaged over the levels of: curva, gen
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## 2 - 5 -0.390 0.0335 36 -11.657 <.0001
## 2 - 6 -0.581 0.0335 36 -17.353 <.0001
## 5 - 6 -0.191 0.0335 36 -5.696 <.0001
##
## Results are averaged over the levels of: curva, gen
## P value adjustment: tukey method for comparing a family of 3 estimates
emm_diam2_curva <- emmeans(res.aov.error, pairwise ~ diam2 | curva)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_diam2_curva
## $emmeans
## curva = T3:
## diam2 emmean SE df lower.CL upper.CL
## 2 0.683 0.046 49.8 0.591 0.775
## 5 1.144 0.046 49.8 1.052 1.236
## 6 1.417 0.046 49.8 1.324 1.509
##
## curva = T1:
## diam2 emmean SE df lower.CL upper.CL
## 2 0.872 0.046 49.8 0.779 0.964
## 5 1.242 0.046 49.8 1.149 1.334
## 6 1.341 0.046 49.8 1.249 1.434
##
## curva = T2:
## diam2 emmean SE df lower.CL upper.CL
## 2 0.704 0.046 49.8 0.611 0.796
## 5 1.043 0.046 49.8 0.950 1.135
## 6 1.242 0.046 49.8 1.150 1.334
##
## Results are averaged over the levels of: gen
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## curva = T3:
## contrast estimate SE df t.ratio p.value
## 2 - 5 -0.4611 0.0579 36 -7.957 <.0001
## 2 - 6 -0.7338 0.0579 36 -12.664 <.0001
## 5 - 6 -0.2727 0.0579 36 -4.706 0.0001
##
## curva = T1:
## contrast estimate SE df t.ratio p.value
## 2 - 5 -0.3700 0.0579 36 -6.384 <.0001
## 2 - 6 -0.4696 0.0579 36 -8.104 <.0001
## 5 - 6 -0.0996 0.0579 36 -1.720 0.2118
##
## curva = T2:
## contrast estimate SE df t.ratio p.value
## 2 - 5 -0.3390 0.0579 36 -5.849 <.0001
## 2 - 6 -0.5383 0.0579 36 -9.290 <.0001
## 5 - 6 -0.1994 0.0579 36 -3.441 0.0041
##
## Results are averaged over the levels of: gen
## P value adjustment: tukey method for comparing a family of 3 estimates
## PI-ITF
##Summary statistics
summ<-datos %>%
group_by(curva, gen) %>%
get_summary_stats(pi.tf, type = "mean_sd")
summ %>% as_tibble() %>% print(n=Inf)
## # A tibble: 9 × 6
## curva gen variable n mean sd
## <fct> <fct> <chr> <dbl> <dbl> <dbl>
## 1 T3 CCN51 pi.tf 9 1.88 1.07
## 2 T3 ICS95 pi.tf 9 1.44 0.785
## 3 T3 TCS01 pi.tf 9 1.95 0.849
## 4 T1 CCN51 pi.tf 9 1.67 1.17
## 5 T1 ICS95 pi.tf 9 0.517 0.227
## 6 T1 TCS01 pi.tf 9 0.997 0.548
## 7 T2 CCN51 pi.tf 9 0.801 0.577
## 8 T2 ICS95 pi.tf 9 0.433 0.203
## 9 T2 TCS01 pi.tf 9 1.04 0.693
##Visualization
bxp <- ggboxplot(
datos, x = "curva", y = "pi.tf",
palette = "jco",
color = "gen", xlab = "Treatment", ylab = "PI-ITF", legend.title = "Genotype"
)
bxp
##Check assumptions
##Outliers
datos %>%
group_by(curva, gen) %>%
identify_outliers(pi.tf)
## # A tibble: 4 × 21
## curva gen ids trat muestra id diam2 cd.testa cd.grano pdcd.grano
## <fct> <fct> <int> <chr> <fct> <fct> <fct> <dbl> <dbl> <dbl>
## 1 T3 TCS01 32 T3 25 8 6 8.89 7.02 -0.260
## 2 T1 CCN51 39 T1 14 10 5 12.0 7.68 -0.232
## 3 T1 CCN51 40 T1 14 10 6 13.7 7.64 -0.235
## 4 T2 ICS95 88 T2 14 22 6 11.3 10.2 -0.183
## # ℹ 11 more variables: picd.testa <dbl>, pdph.grano <dbl>, piph.testa <dbl>,
## # ph.testa <dbl>, acidez.testa <dbl>, ph.grano <dbl>, acidez.grano <dbl>,
## # tf <dbl>, pi.tf <dbl>, is.outlier <lgl>, is.extreme <lgl>
##Normality assumption
##Compute Shapiro-Wilk test for each combinations of factor levels:
norm<-datos %>%
group_by(curva, gen) %>%
shapiro_test(pi.tf)
norm %>% as_tibble() %>% print(n=Inf)
## # A tibble: 9 × 5
## curva gen variable statistic p
## <fct> <fct> <chr> <dbl> <dbl>
## 1 T3 CCN51 pi.tf 0.948 0.672
## 2 T3 ICS95 pi.tf 0.943 0.609
## 3 T3 TCS01 pi.tf 0.968 0.876
## 4 T1 CCN51 pi.tf 0.909 0.311
## 5 T1 ICS95 pi.tf 0.910 0.317
## 6 T1 TCS01 pi.tf 0.940 0.581
## 7 T2 CCN51 pi.tf 0.842 0.0613
## 8 T2 ICS95 pi.tf 0.952 0.707
## 9 T2 TCS01 pi.tf 0.925 0.433
##Create QQ plot for each cell of design:
ggqqplot(datos, "pi.tf", ggtheme = theme_bw()) +
facet_grid(curva~ gen, labeller = "label_both")
## Warning: The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
##Homogneity of variance assumption
##Compute the Levene’s test at each level of the within-subjects factor, here time variable:
lev<-datos %>%
group_by(curva) %>%
levene_test(pi.tf ~ gen)
lev %>% as_tibble() %>% print(n=Inf)
## # A tibble: 3 × 5
## curva df1 df2 statistic p
## <fct> <int> <int> <dbl> <dbl>
## 1 T3 2 24 0.633 0.540
## 2 T1 2 24 3.57 0.0438
## 3 T2 2 24 2.58 0.0966
#Table by error
res.aov.error <- aov(pi.tf ~ diam2*curva*gen + Error(id/diam2), datos)
res.aov.error
##
## Call:
## aov(formula = pi.tf ~ diam2 * curva * gen + Error(id/diam2),
## data = datos)
##
## Grand Mean: 1.191778
##
## Stratum 1: id
##
## Terms:
## curva gen curva:gen Residuals
## Sum of Squares 14.036789 6.568553 2.580395 10.130900
## Deg. of Freedom 2 2 4 18
##
## Residual standard error: 0.7502185
## 16 out of 24 effects not estimable
## Estimated effects may be unbalanced
##
## Stratum 2: id:diam2
##
## Terms:
## diam2 diam2:curva diam2:gen diam2:curva:gen Residuals
## Sum of Squares 20.804200 2.729280 2.109180 0.925020 3.765823
## Deg. of Freedom 2 4 4 8 36
##
## Residual standard error: 0.3234288
## Estimated effects may be unbalanced
summary(res.aov.error)
##
## Error: id
## Df Sum Sq Mean Sq F value Pr(>F)
## curva 2 14.037 7.018 12.470 0.0004 ***
## gen 2 6.569 3.284 5.835 0.0111 *
## curva:gen 4 2.580 0.645 1.146 0.3668
## Residuals 18 10.131 0.563
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Error: id:diam2
## Df Sum Sq Mean Sq F value Pr(>F)
## diam2 2 20.804 10.402 99.441 2.18e-15 ***
## diam2:curva 4 2.729 0.682 6.523 0.000469 ***
## diam2:gen 4 2.109 0.527 5.041 0.002490 **
## diam2:curva:gen 8 0.925 0.116 1.105 0.382569
## Residuals 36 3.766 0.105
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## Emmeans
emmip(res.aov.error, gen ~ diam2 | curva)
## Note: re-fitting model with sum-to-zero contrasts
emm_curva <- emmeans(res.aov.error, pairwise ~ curva)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_curva
## $emmeans
## curva emmean SE df lower.CL upper.CL
## T3 1.754 0.144 18 1.450 2.06
## T1 1.062 0.144 18 0.759 1.37
## T2 0.759 0.144 18 0.456 1.06
##
## Results are averaged over the levels of: diam2, gen
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## T3 - T1 0.691 0.204 18 3.386 0.0088
## T3 - T2 0.995 0.204 18 4.872 0.0003
## T1 - T2 0.303 0.204 18 1.486 0.3207
##
## Results are averaged over the levels of: diam2, gen
## P value adjustment: tukey method for comparing a family of 3 estimates
emm_gen_curva <- emmeans(res.aov.error, pairwise ~ gen | curva)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_gen_curva
## $emmeans
## curva = T3:
## gen emmean SE df lower.CL upper.CL
## CCN51 1.880 0.25 18 1.3548 2.406
## ICS95 1.436 0.25 18 0.9102 1.961
## TCS01 1.946 0.25 18 1.4203 2.471
##
## curva = T1:
## gen emmean SE df lower.CL upper.CL
## CCN51 1.673 0.25 18 1.1476 2.198
## ICS95 0.517 0.25 18 -0.0081 1.043
## TCS01 0.997 0.25 18 0.4718 1.523
##
## curva = T2:
## gen emmean SE df lower.CL upper.CL
## CCN51 0.801 0.25 18 0.2759 1.327
## ICS95 0.433 0.25 18 -0.0921 0.959
## TCS01 1.043 0.25 18 0.5172 1.568
##
## Results are averaged over the levels of: diam2
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## curva = T3:
## contrast estimate SE df t.ratio p.value
## CCN51 - ICS95 0.4446 0.354 18 1.257 0.4364
## CCN51 - TCS01 -0.0654 0.354 18 -0.185 0.9813
## ICS95 - TCS01 -0.5101 0.354 18 -1.442 0.3412
##
## curva = T1:
## contrast estimate SE df t.ratio p.value
## CCN51 - ICS95 1.1557 0.354 18 3.268 0.0113
## CCN51 - TCS01 0.6758 0.354 18 1.911 0.1644
## ICS95 - TCS01 -0.4799 0.354 18 -1.357 0.3835
##
## curva = T2:
## contrast estimate SE df t.ratio p.value
## CCN51 - ICS95 0.3681 0.354 18 1.041 0.5616
## CCN51 - TCS01 -0.2412 0.354 18 -0.682 0.7767
## ICS95 - TCS01 -0.6093 0.354 18 -1.723 0.2241
##
## Results are averaged over the levels of: diam2
## P value adjustment: tukey method for comparing a family of 3 estimates
emm_gen_diam2 <- emmeans(res.aov.error, pairwise ~ curva*gen)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_gen_diam2
## $emmeans
## curva gen emmean SE df lower.CL upper.CL
## T3 CCN51 1.880 0.25 18 1.3548 2.406
## T1 CCN51 1.673 0.25 18 1.1476 2.198
## T2 CCN51 0.801 0.25 18 0.2759 1.327
## T3 ICS95 1.436 0.25 18 0.9102 1.961
## T1 ICS95 0.517 0.25 18 -0.0081 1.043
## T2 ICS95 0.433 0.25 18 -0.0921 0.959
## T3 TCS01 1.946 0.25 18 1.4203 2.471
## T1 TCS01 0.997 0.25 18 0.4718 1.523
## T2 TCS01 1.043 0.25 18 0.5172 1.568
##
## Results are averaged over the levels of: diam2
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## T3 CCN51 - T1 CCN51 0.2073 0.354 18 0.586 0.9995
## T3 CCN51 - T2 CCN51 1.0789 0.354 18 3.051 0.1175
## T3 CCN51 - T3 ICS95 0.4446 0.354 18 1.257 0.9314
## T3 CCN51 - T1 ICS95 1.3629 0.354 18 3.854 0.0249
## T3 CCN51 - T2 ICS95 1.4470 0.354 18 4.091 0.0154
## T3 CCN51 - T3 TCS01 -0.0654 0.354 18 -0.185 1.0000
## T3 CCN51 - T1 TCS01 0.8830 0.354 18 2.497 0.2952
## T3 CCN51 - T2 TCS01 0.8377 0.354 18 2.369 0.3555
## T1 CCN51 - T2 CCN51 0.8716 0.354 18 2.465 0.3097
## T1 CCN51 - T3 ICS95 0.2374 0.354 18 0.671 0.9986
## T1 CCN51 - T1 ICS95 1.1557 0.354 18 3.268 0.0787
## T1 CCN51 - T2 ICS95 1.2397 0.354 18 3.505 0.0498
## T1 CCN51 - T3 TCS01 -0.2727 0.354 18 -0.771 0.9964
## T1 CCN51 - T1 TCS01 0.6758 0.354 18 1.911 0.6156
## T1 CCN51 - T2 TCS01 0.6304 0.354 18 1.783 0.6921
## T2 CCN51 - T3 ICS95 -0.6343 0.354 18 -1.793 0.6857
## T2 CCN51 - T1 ICS95 0.2840 0.354 18 0.803 0.9953
## T2 CCN51 - T2 ICS95 0.3681 0.354 18 1.041 0.9759
## T2 CCN51 - T3 TCS01 -1.1444 0.354 18 -3.236 0.0836
## T2 CCN51 - T1 TCS01 -0.1959 0.354 18 -0.554 0.9997
## T2 CCN51 - T2 TCS01 -0.2412 0.354 18 -0.682 0.9985
## T3 ICS95 - T1 ICS95 0.9183 0.354 18 2.597 0.2535
## T3 ICS95 - T2 ICS95 1.0023 0.354 18 2.834 0.1718
## T3 ICS95 - T3 TCS01 -0.5101 0.354 18 -1.442 0.8668
## T3 ICS95 - T1 TCS01 0.4384 0.354 18 1.240 0.9363
## T3 ICS95 - T2 TCS01 0.3930 0.354 18 1.111 0.9648
## T1 ICS95 - T2 ICS95 0.0840 0.354 18 0.238 1.0000
## T1 ICS95 - T3 TCS01 -1.4284 0.354 18 -4.039 0.0171
## T1 ICS95 - T1 TCS01 -0.4799 0.354 18 -1.357 0.8997
## T1 ICS95 - T2 TCS01 -0.5253 0.354 18 -1.485 0.8484
## T2 ICS95 - T3 TCS01 -1.5124 0.354 18 -4.277 0.0105
## T2 ICS95 - T1 TCS01 -0.5640 0.354 18 -1.595 0.7959
## T2 ICS95 - T2 TCS01 -0.6093 0.354 18 -1.723 0.7265
## T3 TCS01 - T1 TCS01 0.9485 0.354 18 2.682 0.2213
## T3 TCS01 - T2 TCS01 0.9031 0.354 18 2.554 0.2709
## T1 TCS01 - T2 TCS01 -0.0453 0.354 18 -0.128 1.0000
##
## Results are averaged over the levels of: diam2
## P value adjustment: tukey method for comparing a family of 9 estimates
emm_gen_diam2_trend <- emmeans(res.aov.error, pairwise ~ gen)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_gen_diam2_trend
## $emmeans
## gen emmean SE df lower.CL upper.CL
## CCN51 1.451 0.144 18 1.148 1.75
## ICS95 0.795 0.144 18 0.492 1.10
## TCS01 1.328 0.144 18 1.025 1.63
##
## Results are averaged over the levels of: diam2, curva
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## CCN51 - ICS95 0.656 0.204 18 3.213 0.0127
## CCN51 - TCS01 0.123 0.204 18 0.603 0.8205
## ICS95 - TCS01 -0.533 0.204 18 -2.611 0.0445
##
## Results are averaged over the levels of: diam2, curva
## P value adjustment: tukey method for comparing a family of 3 estimates
emm_diam2 <- emmeans(res.aov.error, pairwise ~ diam2)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_diam2
## $emmeans
## diam2 emmean SE df lower.CL upper.CL
## 2 0.52 0.0976 31.7 0.322 0.719
## 5 1.31 0.0976 31.7 1.111 1.509
## 6 1.74 0.0976 31.7 1.546 1.944
##
## Results are averaged over the levels of: curva, gen
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## 2 - 5 -0.790 0.088 36 -8.970 <.0001
## 2 - 6 -1.224 0.088 36 -13.909 <.0001
## 5 - 6 -0.435 0.088 36 -4.940 0.0001
##
## Results are averaged over the levels of: curva, gen
## P value adjustment: tukey method for comparing a family of 3 estimates
emm_diam2_curva <- emmeans(res.aov.error, pairwise ~ diam2 | curva)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_diam2_curva
## $emmeans
## curva = T3:
## diam2 emmean SE df lower.CL upper.CL
## 2 0.755 0.169 31.7 0.4104 1.100
## 5 1.903 0.169 31.7 1.5586 2.248
## 6 2.603 0.169 31.7 2.2587 2.948
##
## curva = T1:
## diam2 emmean SE df lower.CL upper.CL
## 2 0.554 0.169 31.7 0.2097 0.899
## 5 1.213 0.169 31.7 0.8685 1.558
## 6 1.420 0.169 31.7 1.0755 1.765
##
## curva = T2:
## diam2 emmean SE df lower.CL upper.CL
## 2 0.252 0.169 31.7 -0.0925 0.597
## 5 0.814 0.169 31.7 0.4692 1.158
## 6 1.211 0.169 31.7 0.8666 1.556
##
## Results are averaged over the levels of: gen
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## curva = T3:
## contrast estimate SE df t.ratio p.value
## 2 - 5 -1.148 0.152 36 -7.531 <.0001
## 2 - 6 -1.848 0.152 36 -12.123 <.0001
## 5 - 6 -0.700 0.152 36 -4.592 0.0002
##
## curva = T1:
## contrast estimate SE df t.ratio p.value
## 2 - 5 -0.659 0.152 36 -4.321 0.0003
## 2 - 6 -0.866 0.152 36 -5.678 <.0001
## 5 - 6 -0.207 0.152 36 -1.357 0.3736
##
## curva = T2:
## contrast estimate SE df t.ratio p.value
## 2 - 5 -0.562 0.152 36 -3.684 0.0021
## 2 - 6 -0.959 0.152 36 -6.290 <.0001
## 5 - 6 -0.397 0.152 36 -2.606 0.0345
##
## Results are averaged over the levels of: gen
## P value adjustment: tukey method for comparing a family of 3 estimates
## PD-Cd
##Summary statistics
summ<-datos %>%
group_by(curva, gen) %>%
get_summary_stats(pdcd.grano, type = "mean_sd")
summ %>% as_tibble() %>% print(n=Inf)
## # A tibble: 9 × 6
## curva gen variable n mean sd
## <fct> <fct> <chr> <dbl> <dbl> <dbl>
## 1 T3 CCN51 pdcd.grano 9 -0.106 0.074
## 2 T3 ICS95 pdcd.grano 9 -0.074 0.066
## 3 T3 TCS01 pdcd.grano 9 -0.192 0.104
## 4 T1 CCN51 pdcd.grano 9 -0.162 0.062
## 5 T1 ICS95 pdcd.grano 9 -0.08 0.05
## 6 T1 TCS01 pdcd.grano 9 -0.134 0.114
## 7 T2 CCN51 pdcd.grano 9 -0.121 0.089
## 8 T2 ICS95 pdcd.grano 9 -0.074 0.05
## 9 T2 TCS01 pdcd.grano 9 -0.122 0.105
##Visualization
bxp <- ggboxplot(
datos, x = "curva", y = "pdcd.grano",
palette = "jco",
color = "gen", xlab = "Treatment", ylab = "PD-Cd", legend.title = "Genotype"
)
bxp
##Check assumptions
##Outliers
datos %>%
group_by(curva, gen) %>%
identify_outliers(pdcd.grano)
## # A tibble: 3 × 21
## curva gen ids trat muestra id diam2 cd.testa cd.grano pdcd.grano
## <fct> <fct> <int> <chr> <fct> <fct> <fct> <dbl> <dbl> <dbl>
## 1 T3 ICS95 24 T3 36 6 6 14.4 9.48 -0.197
## 2 T1 ICS95 60 T1 36 15 6 8.69 8.06 -0.185
## 3 T2 ICS95 88 T2 14 22 6 11.3 10.2 -0.183
## # ℹ 11 more variables: picd.testa <dbl>, pdph.grano <dbl>, piph.testa <dbl>,
## # ph.testa <dbl>, acidez.testa <dbl>, ph.grano <dbl>, acidez.grano <dbl>,
## # tf <dbl>, pi.tf <dbl>, is.outlier <lgl>, is.extreme <lgl>
##Normality assumption
##Compute Shapiro-Wilk test for each combinations of factor levels:
norm<-datos %>%
group_by(curva, gen) %>%
shapiro_test(pdcd.grano)
norm %>% as_tibble() %>% print(n=Inf)
## # A tibble: 9 × 5
## curva gen variable statistic p
## <fct> <fct> <chr> <dbl> <dbl>
## 1 T3 CCN51 pdcd.grano 0.907 0.297
## 2 T3 ICS95 pdcd.grano 0.852 0.0783
## 3 T3 TCS01 pdcd.grano 0.966 0.856
## 4 T1 CCN51 pdcd.grano 0.945 0.632
## 5 T1 ICS95 pdcd.grano 0.905 0.282
## 6 T1 TCS01 pdcd.grano 0.893 0.215
## 7 T2 CCN51 pdcd.grano 0.917 0.369
## 8 T2 ICS95 pdcd.grano 0.916 0.360
## 9 T2 TCS01 pdcd.grano 0.844 0.0642
##Create QQ plot for each cell of design:
ggqqplot(datos, "pdcd.grano", ggtheme = theme_bw()) +
facet_grid(curva~ gen, labeller = "label_both")
## Warning: The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
##Homogneity of variance assumption
##Compute the Levene’s test at each level of the within-subjects factor, here time variable:
lev<-datos %>%
group_by(curva) %>%
levene_test(pdcd.grano ~ gen)
lev %>% as_tibble() %>% print(n=Inf)
## # A tibble: 3 × 5
## curva df1 df2 statistic p
## <fct> <int> <int> <dbl> <dbl>
## 1 T3 2 24 0.868 0.432
## 2 T1 2 24 2.74 0.0845
## 3 T2 2 24 2.03 0.154
##Computation
res.aov <- anova_test(
data = datos, dv = pdcd.grano, wid = id,
within = diam2, between = c(curva, gen)
)
res.aov
## ANOVA Table (type II tests)
##
## $ANOVA
## Effect DFn DFd F p p<.05 ges
## 1 curva 2 18 0.241 7.89e-01 0.023
## 2 gen 2 18 2.918 8.00e-02 0.221
## 3 diam2 2 36 90.975 8.38e-15 * 0.384
## 4 curva:gen 4 18 0.634 6.45e-01 0.110
## 5 curva:diam2 4 36 5.825 1.00e-03 * 0.074
## 6 gen:diam2 4 36 1.669 1.79e-01 0.022
## 7 curva:gen:diam2 8 36 2.783 1.70e-02 * 0.071
##
## $`Mauchly's Test for Sphericity`
## Effect W p p<.05
## 1 diam2 0.867 0.297
## 2 curva:diam2 0.867 0.297
## 3 gen:diam2 0.867 0.297
## 4 curva:gen:diam2 0.867 0.297
##
## $`Sphericity Corrections`
## Effect GGe DF[GG] p[GG] p[GG]<.05 HFe DF[HF]
## 1 diam2 0.883 1.77, 31.77 2.56e-13 * 0.971 1.94, 34.97
## 2 curva:diam2 0.883 3.53, 31.77 2.00e-03 * 0.971 3.89, 34.97
## 3 gen:diam2 0.883 3.53, 31.77 1.87e-01 0.971 3.89, 34.97
## 4 curva:gen:diam2 0.883 7.06, 31.77 2.20e-02 * 0.971 7.77, 34.97
## p[HF] p[HF]<.05
## 1 1.93e-14 *
## 2 1.00e-03 *
## 3 1.81e-01
## 4 1.80e-02 *
get_anova_table(res.aov)
## ANOVA Table (type II tests)
##
## Effect DFn DFd F p p<.05 ges
## 1 curva 2 18 0.241 7.89e-01 0.023
## 2 gen 2 18 2.918 8.00e-02 0.221
## 3 diam2 2 36 90.975 8.38e-15 * 0.384
## 4 curva:gen 4 18 0.634 6.45e-01 0.110
## 5 curva:diam2 4 36 5.825 1.00e-03 * 0.074
## 6 gen:diam2 4 36 1.669 1.79e-01 0.022
## 7 curva:gen:diam2 8 36 2.783 1.70e-02 * 0.071
#Table by error
res.aov.error <- aov(pdcd.grano ~ diam2*curva*gen + Error(id/diam2), datos)
res.aov.error
##
## Call:
## aov(formula = pdcd.grano ~ diam2 * curva * gen + Error(id/diam2),
## data = datos)
##
## Grand Mean: -0.1183527
##
## Stratum 1: id
##
## Terms:
## curva gen curva:gen Residuals
## Sum of Squares 0.00644012 0.07806532 0.03390014 0.24078781
## Deg. of Freedom 2 2 4 18
##
## Residual standard error: 0.1156594
## 16 out of 24 effects not estimable
## Estimated effects may be unbalanced
##
## Stratum 2: id:diam2
##
## Terms:
## diam2 diam2:curva diam2:gen diam2:curva:gen Residuals
## Sum of Squares 0.17081800 0.02187461 0.00626723 0.02090317 0.03379761
## Deg. of Freedom 2 4 4 8 36
##
## Residual standard error: 0.03064021
## Estimated effects may be unbalanced
summary(res.aov.error)
##
## Error: id
## Df Sum Sq Mean Sq F value Pr(>F)
## curva 2 0.00644 0.00322 0.241 0.7886
## gen 2 0.07807 0.03903 2.918 0.0799 .
## curva:gen 4 0.03390 0.00848 0.634 0.6450
## Residuals 18 0.24079 0.01338
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Error: id:diam2
## Df Sum Sq Mean Sq F value Pr(>F)
## diam2 2 0.17082 0.08541 90.975 8.38e-15 ***
## diam2:curva 4 0.02187 0.00547 5.825 0.00101 **
## diam2:gen 4 0.00627 0.00157 1.669 0.17856
## diam2:curva:gen 8 0.02090 0.00261 2.783 0.01661 *
## Residuals 36 0.03380 0.00094
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## Emmeans
emmip(res.aov.error, gen ~ diam2 | curva)
## Note: re-fitting model with sum-to-zero contrasts
emm_curva <- emmeans(res.aov.error, pairwise ~ curva)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_curva
## $emmeans
## curva emmean SE df lower.CL upper.CL
## T3 -0.124 0.0223 18 -0.171 -0.0772
## T1 -0.125 0.0223 18 -0.172 -0.0786
## T2 -0.106 0.0223 18 -0.153 -0.0590
##
## Results are averaged over the levels of: diam2, gen
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## T3 - T1 0.00137 0.0315 18 0.044 0.9990
## T3 - T2 -0.01819 0.0315 18 -0.578 0.8335
## T1 - T2 -0.01956 0.0315 18 -0.622 0.8103
##
## Results are averaged over the levels of: diam2, gen
## P value adjustment: tukey method for comparing a family of 3 estimates
emm_gen_curva <- emmeans(res.aov.error, pairwise ~ gen | curva)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_gen_curva
## $emmeans
## curva = T3:
## gen emmean SE df lower.CL upper.CL
## CCN51 -0.1060 0.0386 18 -0.187 -0.02496
## ICS95 -0.0741 0.0386 18 -0.155 0.00693
## TCS01 -0.1918 0.0386 18 -0.273 -0.11085
##
## curva = T1:
## gen emmean SE df lower.CL upper.CL
## CCN51 -0.1621 0.0386 18 -0.243 -0.08108
## ICS95 -0.0797 0.0386 18 -0.161 0.00132
## TCS01 -0.1343 0.0386 18 -0.215 -0.05325
##
## curva = T2:
## gen emmean SE df lower.CL upper.CL
## CCN51 -0.1211 0.0386 18 -0.202 -0.04008
## ICS95 -0.0741 0.0386 18 -0.155 0.00689
## TCS01 -0.1221 0.0386 18 -0.203 -0.04113
##
## Results are averaged over the levels of: diam2
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## curva = T3:
## contrast estimate SE df t.ratio p.value
## CCN51 - ICS95 -0.03189 0.0545 18 -0.585 0.8299
## CCN51 - TCS01 0.08589 0.0545 18 1.575 0.2814
## ICS95 - TCS01 0.11778 0.0545 18 2.160 0.1057
##
## curva = T1:
## contrast estimate SE df t.ratio p.value
## CCN51 - ICS95 -0.08240 0.0545 18 -1.511 0.3093
## CCN51 - TCS01 -0.02782 0.0545 18 -0.510 0.8673
## ICS95 - TCS01 0.05458 0.0545 18 1.001 0.5857
##
## curva = T2:
## contrast estimate SE df t.ratio p.value
## CCN51 - ICS95 -0.04697 0.0545 18 -0.862 0.6706
## CCN51 - TCS01 0.00105 0.0545 18 0.019 0.9998
## ICS95 - TCS01 0.04802 0.0545 18 0.881 0.6589
##
## Results are averaged over the levels of: diam2
## P value adjustment: tukey method for comparing a family of 3 estimates
emm_gen_diam2 <- emmeans(res.aov.error, pairwise ~ diam2 | curva*gen)
## Note: re-fitting model with sum-to-zero contrasts
emm_gen_diam2
## $emmeans
## curva = T3, gen = CCN51:
## diam2 emmean SE df lower.CL upper.CL
## 2 -0.0580 0.0412 23.2 -0.143 0.027121
## 5 -0.1071 0.0412 23.2 -0.192 -0.021958
## 6 -0.1528 0.0412 23.2 -0.238 -0.067644
##
## curva = T1, gen = CCN51:
## diam2 emmean SE df lower.CL upper.CL
## 2 -0.1213 0.0412 23.2 -0.206 -0.036186
## 5 -0.1721 0.0412 23.2 -0.257 -0.086960
## 6 -0.1928 0.0412 23.2 -0.278 -0.107680
##
## curva = T2, gen = CCN51:
## diam2 emmean SE df lower.CL upper.CL
## 2 -0.0399 0.0412 23.2 -0.125 0.045262
## 5 -0.1073 0.0412 23.2 -0.192 -0.022143
## 6 -0.2161 0.0412 23.2 -0.301 -0.130954
##
## curva = T3, gen = ICS95:
## diam2 emmean SE df lower.CL upper.CL
## 2 -0.0231 0.0412 23.2 -0.108 0.062048
## 5 -0.0684 0.0412 23.2 -0.154 0.016700
## 6 -0.1307 0.0412 23.2 -0.216 -0.045567
##
## curva = T1, gen = ICS95:
## diam2 emmean SE df lower.CL upper.CL
## 2 -0.0392 0.0412 23.2 -0.124 0.045882
## 5 -0.0714 0.0412 23.2 -0.157 0.013683
## 6 -0.1283 0.0412 23.2 -0.213 -0.043192
##
## curva = T2, gen = ICS95:
## diam2 emmean SE df lower.CL upper.CL
## 2 -0.0501 0.0412 23.2 -0.135 0.035048
## 5 -0.0597 0.0412 23.2 -0.145 0.025412
## 6 -0.1125 0.0412 23.2 -0.198 -0.027379
##
## curva = T3, gen = TCS01:
## diam2 emmean SE df lower.CL upper.CL
## 2 -0.1271 0.0412 23.2 -0.212 -0.041989
## 5 -0.1886 0.0412 23.2 -0.274 -0.103512
## 6 -0.2598 0.0412 23.2 -0.345 -0.174653
##
## curva = T1, gen = TCS01:
## diam2 emmean SE df lower.CL upper.CL
## 2 -0.1121 0.0412 23.2 -0.197 -0.026965
## 5 -0.1369 0.0412 23.2 -0.222 -0.051787
## 6 -0.1537 0.0412 23.2 -0.239 -0.068609
##
## curva = T2, gen = TCS01:
## diam2 emmean SE df lower.CL upper.CL
## 2 -0.0261 0.0412 23.2 -0.111 0.059008
## 5 -0.0847 0.0412 23.2 -0.170 0.000429
## 6 -0.2555 0.0412 23.2 -0.341 -0.170417
##
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## curva = T3, gen = CCN51:
## contrast estimate SE df t.ratio p.value
## 2 - 5 0.04908 0.025 36 1.962 0.1364
## 2 - 6 0.09477 0.025 36 3.788 0.0016
## 5 - 6 0.04569 0.025 36 1.826 0.1755
##
## curva = T1, gen = CCN51:
## contrast estimate SE df t.ratio p.value
## 2 - 5 0.05077 0.025 36 2.030 0.1197
## 2 - 6 0.07149 0.025 36 2.858 0.0189
## 5 - 6 0.02072 0.025 36 0.828 0.6882
##
## curva = T2, gen = CCN51:
## contrast estimate SE df t.ratio p.value
## 2 - 5 0.06740 0.025 36 2.694 0.0280
## 2 - 6 0.17622 0.025 36 7.044 <.0001
## 5 - 6 0.10881 0.025 36 4.349 0.0003
##
## curva = T3, gen = ICS95:
## contrast estimate SE df t.ratio p.value
## 2 - 5 0.04535 0.025 36 1.813 0.1799
## 2 - 6 0.10762 0.025 36 4.302 0.0004
## 5 - 6 0.06227 0.025 36 2.489 0.0452
##
## curva = T1, gen = ICS95:
## contrast estimate SE df t.ratio p.value
## 2 - 5 0.03220 0.025 36 1.287 0.4115
## 2 - 6 0.08907 0.025 36 3.560 0.0030
## 5 - 6 0.05688 0.025 36 2.273 0.0726
##
## curva = T2, gen = ICS95:
## contrast estimate SE df t.ratio p.value
## 2 - 5 0.00964 0.025 36 0.385 0.9216
## 2 - 6 0.06243 0.025 36 2.495 0.0446
## 5 - 6 0.05279 0.025 36 2.110 0.1019
##
## curva = T3, gen = TCS01:
## contrast estimate SE df t.ratio p.value
## 2 - 5 0.06152 0.025 36 2.459 0.0484
## 2 - 6 0.13266 0.025 36 5.303 <.0001
## 5 - 6 0.07114 0.025 36 2.844 0.0195
##
## curva = T1, gen = TCS01:
## contrast estimate SE df t.ratio p.value
## 2 - 5 0.02482 0.025 36 0.992 0.5865
## 2 - 6 0.04164 0.025 36 1.665 0.2326
## 5 - 6 0.01682 0.025 36 0.672 0.7809
##
## curva = T2, gen = TCS01:
## contrast estimate SE df t.ratio p.value
## 2 - 5 0.05858 0.025 36 2.342 0.0627
## 2 - 6 0.22943 0.025 36 9.171 <.0001
## 5 - 6 0.17085 0.025 36 6.829 <.0001
##
## P value adjustment: tukey method for comparing a family of 3 estimates
emm_gen_diam2_trend <- emmeans(res.aov.error, pairwise ~ diam2*gen | curva)
## Note: re-fitting model with sum-to-zero contrasts
emm_gen_diam2_trend
## $emmeans
## curva = T3:
## diam2 gen emmean SE df lower.CL upper.CL
## 2 CCN51 -0.0580 0.0412 23.2 -0.143 0.027121
## 5 CCN51 -0.1071 0.0412 23.2 -0.192 -0.021958
## 6 CCN51 -0.1528 0.0412 23.2 -0.238 -0.067644
## 2 ICS95 -0.0231 0.0412 23.2 -0.108 0.062048
## 5 ICS95 -0.0684 0.0412 23.2 -0.154 0.016700
## 6 ICS95 -0.1307 0.0412 23.2 -0.216 -0.045567
## 2 TCS01 -0.1271 0.0412 23.2 -0.212 -0.041989
## 5 TCS01 -0.1886 0.0412 23.2 -0.274 -0.103512
## 6 TCS01 -0.2598 0.0412 23.2 -0.345 -0.174653
##
## curva = T1:
## diam2 gen emmean SE df lower.CL upper.CL
## 2 CCN51 -0.1213 0.0412 23.2 -0.206 -0.036186
## 5 CCN51 -0.1721 0.0412 23.2 -0.257 -0.086960
## 6 CCN51 -0.1928 0.0412 23.2 -0.278 -0.107680
## 2 ICS95 -0.0392 0.0412 23.2 -0.124 0.045882
## 5 ICS95 -0.0714 0.0412 23.2 -0.157 0.013683
## 6 ICS95 -0.1283 0.0412 23.2 -0.213 -0.043192
## 2 TCS01 -0.1121 0.0412 23.2 -0.197 -0.026965
## 5 TCS01 -0.1369 0.0412 23.2 -0.222 -0.051787
## 6 TCS01 -0.1537 0.0412 23.2 -0.239 -0.068609
##
## curva = T2:
## diam2 gen emmean SE df lower.CL upper.CL
## 2 CCN51 -0.0399 0.0412 23.2 -0.125 0.045262
## 5 CCN51 -0.1073 0.0412 23.2 -0.192 -0.022143
## 6 CCN51 -0.2161 0.0412 23.2 -0.301 -0.130954
## 2 ICS95 -0.0501 0.0412 23.2 -0.135 0.035048
## 5 ICS95 -0.0597 0.0412 23.2 -0.145 0.025412
## 6 ICS95 -0.1125 0.0412 23.2 -0.198 -0.027379
## 2 TCS01 -0.0261 0.0412 23.2 -0.111 0.059008
## 5 TCS01 -0.0847 0.0412 23.2 -0.170 0.000429
## 6 TCS01 -0.2555 0.0412 23.2 -0.341 -0.170417
##
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## curva = T3:
## contrast estimate SE df t.ratio p.value
## 2 CCN51 - 5 CCN51 0.04908 0.0250 36.0 1.962 0.5780
## 2 CCN51 - 6 CCN51 0.09477 0.0250 36.0 3.788 0.0145
## 2 CCN51 - 2 ICS95 -0.03493 0.0582 23.2 -0.600 0.9994
## 2 CCN51 - 5 ICS95 0.01042 0.0582 23.2 0.179 1.0000
## 2 CCN51 - 6 ICS95 0.07269 0.0582 23.2 1.248 0.9363
## 2 CCN51 - 2 TCS01 0.06911 0.0582 23.2 1.187 0.9515
## 2 CCN51 - 5 TCS01 0.13063 0.0582 23.2 2.244 0.4123
## 2 CCN51 - 6 TCS01 0.20177 0.0582 23.2 3.466 0.0444
## 5 CCN51 - 6 CCN51 0.04569 0.0250 36.0 1.826 0.6653
## 5 CCN51 - 2 ICS95 -0.08401 0.0582 23.2 -1.443 0.8695
## 5 CCN51 - 5 ICS95 -0.03866 0.0582 23.2 -0.664 0.9988
## 5 CCN51 - 6 ICS95 0.02361 0.0582 23.2 0.405 1.0000
## 5 CCN51 - 2 TCS01 0.02003 0.0582 23.2 0.344 1.0000
## 5 CCN51 - 5 TCS01 0.08155 0.0582 23.2 1.401 0.8864
## 5 CCN51 - 6 TCS01 0.15269 0.0582 23.2 2.623 0.2303
## 6 CCN51 - 2 ICS95 -0.12969 0.0582 23.2 -2.228 0.4215
## 6 CCN51 - 5 ICS95 -0.08434 0.0582 23.2 -1.449 0.8671
## 6 CCN51 - 6 ICS95 -0.02208 0.0582 23.2 -0.379 1.0000
## 6 CCN51 - 2 TCS01 -0.02565 0.0582 23.2 -0.441 0.9999
## 6 CCN51 - 5 TCS01 0.03587 0.0582 23.2 0.616 0.9993
## 6 CCN51 - 6 TCS01 0.10701 0.0582 23.2 1.838 0.6587
## 2 ICS95 - 5 ICS95 0.04535 0.0250 36.0 1.813 0.6739
## 2 ICS95 - 6 ICS95 0.10762 0.0250 36.0 4.302 0.0035
## 2 ICS95 - 2 TCS01 0.10404 0.0582 23.2 1.787 0.6897
## 2 ICS95 - 5 TCS01 0.16556 0.0582 23.2 2.844 0.1556
## 2 ICS95 - 6 TCS01 0.23670 0.0582 23.2 4.065 0.0116
## 5 ICS95 - 6 ICS95 0.06227 0.0250 36.0 2.489 0.2708
## 5 ICS95 - 2 TCS01 0.05869 0.0582 23.2 1.008 0.9812
## 5 ICS95 - 5 TCS01 0.12021 0.0582 23.2 2.065 0.5181
## 5 ICS95 - 6 TCS01 0.19135 0.0582 23.2 3.287 0.0649
## 6 ICS95 - 2 TCS01 -0.00358 0.0582 23.2 -0.061 1.0000
## 6 ICS95 - 5 TCS01 0.05795 0.0582 23.2 0.995 0.9826
## 6 ICS95 - 6 TCS01 0.12909 0.0582 23.2 2.217 0.4274
## 2 TCS01 - 5 TCS01 0.06152 0.0250 36.0 2.459 0.2849
## 2 TCS01 - 6 TCS01 0.13266 0.0250 36.0 5.303 0.0002
## 5 TCS01 - 6 TCS01 0.07114 0.0250 36.0 2.844 0.1378
##
## curva = T1:
## contrast estimate SE df t.ratio p.value
## 2 CCN51 - 5 CCN51 0.05077 0.0250 36.0 2.030 0.5341
## 2 CCN51 - 6 CCN51 0.07149 0.0250 36.0 2.858 0.1338
## 2 CCN51 - 2 ICS95 -0.08207 0.0582 23.2 -1.410 0.8830
## 2 CCN51 - 5 ICS95 -0.04987 0.0582 23.2 -0.857 0.9933
## 2 CCN51 - 6 ICS95 0.00701 0.0582 23.2 0.120 1.0000
## 2 CCN51 - 2 TCS01 -0.00922 0.0582 23.2 -0.158 1.0000
## 2 CCN51 - 5 TCS01 0.01560 0.0582 23.2 0.268 1.0000
## 2 CCN51 - 6 TCS01 0.03242 0.0582 23.2 0.557 0.9997
## 5 CCN51 - 6 CCN51 0.02072 0.0250 36.0 0.828 0.9952
## 5 CCN51 - 2 ICS95 -0.13284 0.0582 23.2 -2.282 0.3913
## 5 CCN51 - 5 ICS95 -0.10064 0.0582 23.2 -1.729 0.7242
## 5 CCN51 - 6 ICS95 -0.04377 0.0582 23.2 -0.752 0.9972
## 5 CCN51 - 2 TCS01 -0.05999 0.0582 23.2 -1.030 0.9785
## 5 CCN51 - 5 TCS01 -0.03517 0.0582 23.2 -0.604 0.9994
## 5 CCN51 - 6 TCS01 -0.01835 0.0582 23.2 -0.315 1.0000
## 6 CCN51 - 2 ICS95 -0.15356 0.0582 23.2 -2.637 0.2245
## 6 CCN51 - 5 ICS95 -0.12136 0.0582 23.2 -2.084 0.5061
## 6 CCN51 - 6 ICS95 -0.06449 0.0582 23.2 -1.108 0.9671
## 6 CCN51 - 2 TCS01 -0.08072 0.0582 23.2 -1.386 0.8919
## 6 CCN51 - 5 TCS01 -0.05589 0.0582 23.2 -0.960 0.9861
## 6 CCN51 - 6 TCS01 -0.03907 0.0582 23.2 -0.671 0.9987
## 2 ICS95 - 5 ICS95 0.03220 0.0250 36.0 1.287 0.9287
## 2 ICS95 - 6 ICS95 0.08907 0.0250 36.0 3.560 0.0261
## 2 ICS95 - 2 TCS01 0.07285 0.0582 23.2 1.251 0.9356
## 2 ICS95 - 5 TCS01 0.09767 0.0582 23.2 1.677 0.7534
## 2 ICS95 - 6 TCS01 0.11449 0.0582 23.2 1.966 0.5790
## 5 ICS95 - 6 ICS95 0.05688 0.0250 36.0 2.273 0.3837
## 5 ICS95 - 2 TCS01 0.04065 0.0582 23.2 0.698 0.9983
## 5 ICS95 - 5 TCS01 0.06547 0.0582 23.2 1.124 0.9642
## 5 ICS95 - 6 TCS01 0.08229 0.0582 23.2 1.413 0.8815
## 6 ICS95 - 2 TCS01 -0.01623 0.0582 23.2 -0.279 1.0000
## 6 ICS95 - 5 TCS01 0.00860 0.0582 23.2 0.148 1.0000
## 6 ICS95 - 6 TCS01 0.02542 0.0582 23.2 0.437 0.9999
## 2 TCS01 - 5 TCS01 0.02482 0.0250 36.0 0.992 0.9843
## 2 TCS01 - 6 TCS01 0.04164 0.0250 36.0 1.665 0.7630
## 5 TCS01 - 6 TCS01 0.01682 0.0250 36.0 0.672 0.9989
##
## curva = T2:
## contrast estimate SE df t.ratio p.value
## 2 CCN51 - 5 CCN51 0.06740 0.0250 36.0 2.694 0.1858
## 2 CCN51 - 6 CCN51 0.17622 0.0250 36.0 7.044 <.0001
## 2 CCN51 - 2 ICS95 0.01021 0.0582 23.2 0.175 1.0000
## 2 CCN51 - 5 ICS95 0.01985 0.0582 23.2 0.341 1.0000
## 2 CCN51 - 6 ICS95 0.07264 0.0582 23.2 1.248 0.9365
## 2 CCN51 - 2 TCS01 -0.01375 0.0582 23.2 -0.236 1.0000
## 2 CCN51 - 5 TCS01 0.04483 0.0582 23.2 0.770 0.9967
## 2 CCN51 - 6 TCS01 0.21568 0.0582 23.2 3.704 0.0263
## 5 CCN51 - 6 CCN51 0.10881 0.0250 36.0 4.349 0.0031
## 5 CCN51 - 2 ICS95 -0.05719 0.0582 23.2 -0.982 0.9840
## 5 CCN51 - 5 ICS95 -0.04755 0.0582 23.2 -0.817 0.9951
## 5 CCN51 - 6 ICS95 0.00524 0.0582 23.2 0.090 1.0000
## 5 CCN51 - 2 TCS01 -0.08115 0.0582 23.2 -1.394 0.8891
## 5 CCN51 - 5 TCS01 -0.02257 0.0582 23.2 -0.388 1.0000
## 5 CCN51 - 6 TCS01 0.14827 0.0582 23.2 2.547 0.2613
## 6 CCN51 - 2 ICS95 -0.16600 0.0582 23.2 -2.851 0.1534
## 6 CCN51 - 5 ICS95 -0.15637 0.0582 23.2 -2.686 0.2066
## 6 CCN51 - 6 ICS95 -0.10358 0.0582 23.2 -1.779 0.6944
## 6 CCN51 - 2 TCS01 -0.18996 0.0582 23.2 -3.263 0.0682
## 6 CCN51 - 5 TCS01 -0.13138 0.0582 23.2 -2.257 0.4051
## 6 CCN51 - 6 TCS01 0.03946 0.0582 23.2 0.678 0.9986
## 2 ICS95 - 5 ICS95 0.00964 0.0250 36.0 0.385 1.0000
## 2 ICS95 - 6 ICS95 0.06243 0.0250 36.0 2.495 0.2678
## 2 ICS95 - 2 TCS01 -0.02396 0.0582 23.2 -0.412 1.0000
## 2 ICS95 - 5 TCS01 0.03462 0.0582 23.2 0.595 0.9995
## 2 ICS95 - 6 TCS01 0.20547 0.0582 23.2 3.529 0.0387
## 5 ICS95 - 6 ICS95 0.05279 0.0250 36.0 2.110 0.4826
## 5 ICS95 - 2 TCS01 -0.03360 0.0582 23.2 -0.577 0.9996
## 5 ICS95 - 5 TCS01 0.02498 0.0582 23.2 0.429 1.0000
## 5 ICS95 - 6 TCS01 0.19583 0.0582 23.2 3.363 0.0552
## 6 ICS95 - 2 TCS01 -0.08639 0.0582 23.2 -1.484 0.8518
## 6 ICS95 - 5 TCS01 -0.02781 0.0582 23.2 -0.478 0.9999
## 6 ICS95 - 6 TCS01 0.14304 0.0582 23.2 2.457 0.3017
## 2 TCS01 - 5 TCS01 0.05858 0.0250 36.0 2.342 0.3456
## 2 TCS01 - 6 TCS01 0.22943 0.0250 36.0 9.171 <.0001
## 5 TCS01 - 6 TCS01 0.17085 0.0250 36.0 6.829 <.0001
##
## P value adjustment: tukey method for comparing a family of 9 estimates
emm_diam2 <- emmeans(res.aov.error, pairwise ~ diam2)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_diam2
## $emmeans
## diam2 emmean SE df lower.CL upper.CL
## 2 -0.0663 0.0137 23.2 -0.0947 -0.0380
## 5 -0.1107 0.0137 23.2 -0.1391 -0.0823
## 6 -0.1780 0.0137 23.2 -0.2064 -0.1497
##
## Results are averaged over the levels of: curva, gen
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## 2 - 5 0.0444 0.00834 36 5.321 <.0001
## 2 - 6 0.1117 0.00834 36 13.395 <.0001
## 5 - 6 0.0673 0.00834 36 8.074 <.0001
##
## Results are averaged over the levels of: curva, gen
## P value adjustment: tukey method for comparing a family of 3 estimates
emm_diam2_curva <- emmeans(res.aov.error, pairwise ~ diam2 | curva)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_diam2_curva
## $emmeans
## curva = T3:
## diam2 emmean SE df lower.CL upper.CL
## 2 -0.0694 0.0238 23.2 -0.1186 -0.0203
## 5 -0.1214 0.0238 23.2 -0.1705 -0.0722
## 6 -0.1811 0.0238 23.2 -0.2302 -0.1319
##
## curva = T1:
## diam2 emmean SE df lower.CL upper.CL
## 2 -0.0909 0.0238 23.2 -0.1400 -0.0417
## 5 -0.1268 0.0238 23.2 -0.1760 -0.0777
## 6 -0.1583 0.0238 23.2 -0.2074 -0.1091
##
## curva = T2:
## diam2 emmean SE df lower.CL upper.CL
## 2 -0.0387 0.0238 23.2 -0.0878 0.0105
## 5 -0.0839 0.0238 23.2 -0.1330 -0.0347
## 6 -0.1947 0.0238 23.2 -0.2439 -0.1456
##
## Results are averaged over the levels of: gen
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## curva = T3:
## contrast estimate SE df t.ratio p.value
## 2 - 5 0.0520 0.0144 36 3.599 0.0027
## 2 - 6 0.1117 0.0144 36 7.732 <.0001
## 5 - 6 0.0597 0.0144 36 4.133 0.0006
##
## curva = T1:
## contrast estimate SE df t.ratio p.value
## 2 - 5 0.0359 0.0144 36 2.488 0.0453
## 2 - 6 0.0674 0.0144 36 4.667 0.0001
## 5 - 6 0.0315 0.0144 36 2.179 0.0886
##
## curva = T2:
## contrast estimate SE df t.ratio p.value
## 2 - 5 0.0452 0.0144 36 3.130 0.0095
## 2 - 6 0.1560 0.0144 36 10.802 <.0001
## 5 - 6 0.1108 0.0144 36 7.672 <.0001
##
## Results are averaged over the levels of: gen
## P value adjustment: tukey method for comparing a family of 3 estimates
## PI-Cd
##Summary statistics
summ<-datos %>%
group_by(curva, gen) %>%
get_summary_stats(picd.testa, type = "mean_sd")
summ %>% as_tibble() %>% print(n=Inf)
## # A tibble: 9 × 6
## curva gen variable n mean sd
## <fct> <fct> <chr> <dbl> <dbl> <dbl>
## 1 T3 CCN51 picd.testa 9 1.52 0.809
## 2 T3 ICS95 picd.testa 9 1.22 0.578
## 3 T3 TCS01 picd.testa 9 1.34 0.609
## 4 T1 CCN51 picd.testa 9 1.18 0.812
## 5 T1 ICS95 picd.testa 9 0.389 0.166
## 6 T1 TCS01 picd.testa 9 0.686 0.319
## 7 T2 CCN51 picd.testa 9 0.553 0.444
## 8 T2 ICS95 picd.testa 9 0.32 0.134
## 9 T2 TCS01 picd.testa 9 0.733 0.394
##Visualization
bxp <- ggboxplot(
datos, x = "curva", y = "picd.testa",
palette = "jco",
color = "gen", xlab = "Treatment", ylab = "PI-Cd", legend.title = "Genotype"
)
bxp
##Check assumptions
##Outliers
datos %>%
group_by(curva, gen) %>%
identify_outliers(picd.testa)
## # A tibble: 2 × 21
## curva gen ids trat muestra id diam2 cd.testa cd.grano pdcd.grano
## <fct> <fct> <int> <chr> <fct> <fct> <fct> <dbl> <dbl> <dbl>
## 1 T1 CCN51 39 T1 14 10 5 12.0 7.68 -0.232
## 2 T1 CCN51 40 T1 14 10 6 13.7 7.64 -0.235
## # ℹ 11 more variables: picd.testa <dbl>, pdph.grano <dbl>, piph.testa <dbl>,
## # ph.testa <dbl>, acidez.testa <dbl>, ph.grano <dbl>, acidez.grano <dbl>,
## # tf <dbl>, pi.tf <dbl>, is.outlier <lgl>, is.extreme <lgl>
##Normality assumption
##Compute Shapiro-Wilk test for each combinations of factor levels:
norm<-datos %>%
group_by(curva, gen) %>%
shapiro_test(picd.testa)
norm %>% as_tibble() %>% print(n=Inf)
## # A tibble: 9 × 5
## curva gen variable statistic p
## <fct> <fct> <chr> <dbl> <dbl>
## 1 T3 CCN51 picd.testa 0.908 0.299
## 2 T3 ICS95 picd.testa 0.913 0.335
## 3 T3 TCS01 picd.testa 0.929 0.472
## 4 T1 CCN51 picd.testa 0.912 0.329
## 5 T1 ICS95 picd.testa 0.966 0.859
## 6 T1 TCS01 picd.testa 0.957 0.764
## 7 T2 CCN51 picd.testa 0.934 0.518
## 8 T2 ICS95 picd.testa 0.935 0.535
## 9 T2 TCS01 picd.testa 0.906 0.287
##Create QQ plot for each cell of design:
ggqqplot(datos, "picd.testa", ggtheme = theme_bw()) +
facet_grid(curva~ gen, labeller = "label_both")
## Warning: The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
## The following aesthetics were dropped during statistical transformation: sample
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
## the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
## variable into a factor?
##Homogneity of variance assumption
##Compute the Levene’s test at each level of the within-subjects factor, here time variable:
lev<-datos %>%
group_by(curva) %>%
levene_test(picd.testa ~ gen)
lev %>% as_tibble() %>% print(n=Inf)
## # A tibble: 3 × 5
## curva df1 df2 statistic p
## <fct> <int> <int> <dbl> <dbl>
## 1 T3 2 24 0.490 0.619
## 2 T1 2 24 3.52 0.0457
## 3 T2 2 24 2.54 0.0998
#Table by error
res.aov.error <- aov(picd.testa ~ diam2*curva*gen + Error(id/diam2), datos)
res.aov.error
##
## Call:
## aov(formula = picd.testa ~ diam2 * curva * gen + Error(id/diam2),
## data = datos)
##
## Grand Mean: 0.8822763
##
## Stratum 1: id
##
## Terms:
## curva gen curva:gen Residuals
## Sum of Squares 9.844743 2.728207 1.373165 4.481510
## Deg. of Freedom 2 2 4 18
##
## Residual standard error: 0.4989717
## 16 out of 24 effects not estimable
## Estimated effects may be unbalanced
##
## Stratum 2: id:diam2
##
## Terms:
## diam2 diam2:curva diam2:gen diam2:curva:gen Residuals
## Sum of Squares 9.974819 1.679655 1.110548 0.662912 2.229649
## Deg. of Freedom 2 4 4 8 36
##
## Residual standard error: 0.2488668
## Estimated effects may be unbalanced
summary(res.aov.error)
##
## Error: id
## Df Sum Sq Mean Sq F value Pr(>F)
## curva 2 9.845 4.922 19.771 2.87e-05 ***
## gen 2 2.728 1.364 5.479 0.0139 *
## curva:gen 4 1.373 0.343 1.379 0.2807
## Residuals 18 4.482 0.249
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Error: id:diam2
## Df Sum Sq Mean Sq F value Pr(>F)
## diam2 2 9.975 4.987 80.527 5.14e-14 ***
## diam2:curva 4 1.680 0.420 6.780 0.000356 ***
## diam2:gen 4 1.111 0.278 4.483 0.004836 **
## diam2:curva:gen 8 0.663 0.083 1.338 0.256842
## Residuals 36 2.230 0.062
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## Emmeans
emmip(res.aov.error, gen ~ diam2 | curva)
## Note: re-fitting model with sum-to-zero contrasts
emm_curva <- emmeans(res.aov.error, pairwise ~ curva)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_curva
## $emmeans
## curva emmean SE df lower.CL upper.CL
## T3 1.359 0.096 18 1.157 1.561
## T1 0.752 0.096 18 0.551 0.954
## T2 0.535 0.096 18 0.334 0.737
##
## Results are averaged over the levels of: diam2, gen
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## T3 - T1 0.607 0.136 18 4.468 0.0008
## T3 - T2 0.824 0.136 18 6.066 <.0001
## T1 - T2 0.217 0.136 18 1.598 0.2719
##
## Results are averaged over the levels of: diam2, gen
## P value adjustment: tukey method for comparing a family of 3 estimates
emm_gen_curva <- emmeans(res.aov.error, pairwise ~ gen | curva)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_gen_curva
## $emmeans
## curva = T3:
## gen emmean SE df lower.CL upper.CL
## CCN51 1.523 0.166 18 1.1734 1.872
## ICS95 1.215 0.166 18 0.8652 1.564
## TCS01 1.340 0.166 18 0.9905 1.689
##
## curva = T1:
## gen emmean SE df lower.CL upper.CL
## CCN51 1.183 0.166 18 0.8333 1.532
## ICS95 0.389 0.166 18 0.0395 0.738
## TCS01 0.686 0.166 18 0.3361 1.035
##
## curva = T2:
## gen emmean SE df lower.CL upper.CL
## CCN51 0.553 0.166 18 0.2032 0.902
## ICS95 0.320 0.166 18 -0.0294 0.669
## TCS01 0.733 0.166 18 0.3839 1.083
##
## Results are averaged over the levels of: diam2
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## curva = T3:
## contrast estimate SE df t.ratio p.value
## CCN51 - ICS95 0.308 0.235 18 1.310 0.4079
## CCN51 - TCS01 0.183 0.235 18 0.777 0.7213
## ICS95 - TCS01 -0.125 0.235 18 -0.533 0.8564
##
## curva = T1:
## contrast estimate SE df t.ratio p.value
## CCN51 - ICS95 0.794 0.235 18 3.375 0.0090
## CCN51 - TCS01 0.497 0.235 18 2.114 0.1150
## ICS95 - TCS01 -0.297 0.235 18 -1.261 0.4344
##
## curva = T2:
## contrast estimate SE df t.ratio p.value
## CCN51 - ICS95 0.233 0.235 18 0.989 0.5930
## CCN51 - TCS01 -0.181 0.235 18 -0.768 0.7266
## ICS95 - TCS01 -0.413 0.235 18 -1.757 0.2120
##
## Results are averaged over the levels of: diam2
## P value adjustment: tukey method for comparing a family of 3 estimates
emm_gen_diam2 <- emmeans(res.aov.error, pairwise ~ curva*gen)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_gen_diam2
## $emmeans
## curva gen emmean SE df lower.CL upper.CL
## T3 CCN51 1.523 0.166 18 1.1734 1.872
## T1 CCN51 1.183 0.166 18 0.8333 1.532
## T2 CCN51 0.553 0.166 18 0.2032 0.902
## T3 ICS95 1.215 0.166 18 0.8652 1.564
## T1 ICS95 0.389 0.166 18 0.0395 0.738
## T2 ICS95 0.320 0.166 18 -0.0294 0.669
## T3 TCS01 1.340 0.166 18 0.9905 1.689
## T1 TCS01 0.686 0.166 18 0.3361 1.035
## T2 TCS01 0.733 0.166 18 0.3839 1.083
##
## Results are averaged over the levels of: diam2
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## T3 CCN51 - T1 CCN51 0.3401 0.235 18 1.446 0.8654
## T3 CCN51 - T2 CCN51 0.9702 0.235 18 4.125 0.0143
## T3 CCN51 - T3 ICS95 0.3082 0.235 18 1.310 0.9156
## T3 CCN51 - T1 ICS95 1.1339 0.235 18 4.821 0.0034
## T3 CCN51 - T2 ICS95 1.2028 0.235 18 5.114 0.0019
## T3 CCN51 - T3 TCS01 0.1829 0.235 18 0.777 0.9962
## T3 CCN51 - T1 TCS01 0.8373 0.235 18 3.560 0.0448
## T3 CCN51 - T2 TCS01 0.7894 0.235 18 3.356 0.0665
## T1 CCN51 - T2 CCN51 0.6301 0.235 18 2.679 0.2224
## T1 CCN51 - T3 ICS95 -0.0319 0.235 18 -0.136 1.0000
## T1 CCN51 - T1 ICS95 0.7938 0.235 18 3.375 0.0642
## T1 CCN51 - T2 ICS95 0.8627 0.235 18 3.668 0.0362
## T1 CCN51 - T3 TCS01 -0.1572 0.235 18 -0.668 0.9987
## T1 CCN51 - T1 TCS01 0.4972 0.235 18 2.114 0.4941
## T1 CCN51 - T2 TCS01 0.4494 0.235 18 1.910 0.6158
## T2 CCN51 - T3 ICS95 -0.6620 0.235 18 -2.814 0.1777
## T2 CCN51 - T1 ICS95 0.1637 0.235 18 0.696 0.9982
## T2 CCN51 - T2 ICS95 0.2326 0.235 18 0.989 0.9822
## T2 CCN51 - T3 TCS01 -0.7873 0.235 18 -3.347 0.0677
## T2 CCN51 - T1 TCS01 -0.1329 0.235 18 -0.565 0.9996
## T2 CCN51 - T2 TCS01 -0.1807 0.235 18 -0.768 0.9965
## T3 ICS95 - T1 ICS95 0.8257 0.235 18 3.510 0.0494
## T3 ICS95 - T2 ICS95 0.8946 0.235 18 3.803 0.0276
## T3 ICS95 - T3 TCS01 -0.1253 0.235 18 -0.533 0.9997
## T3 ICS95 - T1 TCS01 0.5291 0.235 18 2.249 0.4175
## T3 ICS95 - T2 TCS01 0.4813 0.235 18 2.046 0.5342
## T1 ICS95 - T2 ICS95 0.0689 0.235 18 0.293 1.0000
## T1 ICS95 - T3 TCS01 -0.9510 0.235 18 -4.043 0.0169
## T1 ICS95 - T1 TCS01 -0.2966 0.235 18 -1.261 0.9304
## T1 ICS95 - T2 TCS01 -0.3445 0.235 18 -1.464 0.8575
## T2 ICS95 - T3 TCS01 -1.0199 0.235 18 -4.336 0.0093
## T2 ICS95 - T1 TCS01 -0.3655 0.235 18 -1.554 0.8163
## T2 ICS95 - T2 TCS01 -0.4134 0.235 18 -1.757 0.7067
## T3 TCS01 - T1 TCS01 0.6544 0.235 18 2.782 0.1876
## T3 TCS01 - T2 TCS01 0.6066 0.235 18 2.579 0.2606
## T1 TCS01 - T2 TCS01 -0.0479 0.235 18 -0.203 1.0000
##
## Results are averaged over the levels of: diam2
## P value adjustment: tukey method for comparing a family of 9 estimates
emm_gen_diam2_trend <- emmeans(res.aov.error, pairwise ~ gen)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_gen_diam2_trend
## $emmeans
## gen emmean SE df lower.CL upper.CL
## CCN51 1.086 0.096 18 0.884 1.288
## ICS95 0.641 0.096 18 0.439 0.843
## TCS01 0.920 0.096 18 0.718 1.121
##
## Results are averaged over the levels of: diam2, curva
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## CCN51 - ICS95 0.445 0.136 18 3.276 0.0111
## CCN51 - TCS01 0.166 0.136 18 1.226 0.4539
## ICS95 - TCS01 -0.278 0.136 18 -2.050 0.1290
##
## Results are averaged over the levels of: diam2, curva
## P value adjustment: tukey method for comparing a family of 3 estimates
emm_diam2 <- emmeans(res.aov.error, pairwise ~ diam2)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_diam2
## $emmeans
## diam2 emmean SE df lower.CL upper.CL
## 2 0.399 0.0678 35.9 0.262 0.537
## 5 1.024 0.0678 35.9 0.887 1.162
## 6 1.223 0.0678 35.9 1.086 1.361
##
## Results are averaged over the levels of: curva, gen
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## 2 - 5 -0.625 0.0677 36 -9.223 <.0001
## 2 - 6 -0.824 0.0677 36 -12.161 <.0001
## 5 - 6 -0.199 0.0677 36 -2.937 0.0155
##
## Results are averaged over the levels of: curva, gen
## P value adjustment: tukey method for comparing a family of 3 estimates
emm_diam2_curva <- emmeans(res.aov.error, pairwise ~ diam2 | curva)
## Note: re-fitting model with sum-to-zero contrasts
## NOTE: Results may be misleading due to involvement in interactions
emm_diam2_curva
## $emmeans
## curva = T3:
## diam2 emmean SE df lower.CL upper.CL
## 2 0.612 0.118 35.9 0.3735 0.850
## 5 1.536 0.118 35.9 1.2981 1.775
## 6 1.929 0.118 35.9 1.6908 2.167
##
## curva = T1:
## diam2 emmean SE df lower.CL upper.CL
## 2 0.383 0.118 35.9 0.1448 0.621
## 5 0.883 0.118 35.9 0.6446 1.121
## 6 0.991 0.118 35.9 0.7528 1.229
##
## curva = T2:
## diam2 emmean SE df lower.CL upper.CL
## 2 0.203 0.118 35.9 -0.0349 0.442
## 5 0.653 0.118 35.9 0.4149 0.892
## 6 0.749 0.118 35.9 0.5109 0.988
##
## Results are averaged over the levels of: gen
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95
##
## $contrasts
## curva = T3:
## contrast estimate SE df t.ratio p.value
## 2 - 5 -0.925 0.117 36 -7.881 <.0001
## 2 - 6 -1.317 0.117 36 -11.228 <.0001
## 5 - 6 -0.393 0.117 36 -3.347 0.0053
##
## curva = T1:
## contrast estimate SE df t.ratio p.value
## 2 - 5 -0.500 0.117 36 -4.260 0.0004
## 2 - 6 -0.608 0.117 36 -5.182 <.0001
## 5 - 6 -0.108 0.117 36 -0.922 0.6300
##
## curva = T2:
## contrast estimate SE df t.ratio p.value
## 2 - 5 -0.450 0.117 36 -3.834 0.0014
## 2 - 6 -0.546 0.117 36 -4.652 0.0001
## 5 - 6 -0.096 0.117 36 -0.818 0.6943
##
## Results are averaged over the levels of: gen
## P value adjustment: tukey method for comparing a family of 3 estimates