Ripasudil Study analysis
Juan Steibel
6/14/2021
Analysis of Ripasudil dataset
library(tidyverse)
library(readxl)
library(lubridate)
library(broom)
library(lsmeans)
library(lme4)
library(ggplot2)
library(lmerTest)
library(DT)
setwd("C:/Users/marti/OneDrive/Documents/andras/Ripasudil/")Load data, use graphics and tables to summarize
Read data and show graphics
dts<-read_xlsx("Ripasudil Study Stats Sheet.xlsx",sheet = "Sheet1")
dts<-filter(dts,!is.na(Dog))
colnames(dts)## [1] "Dog" "Age (years)" "Eye"
## [4] "Treatment" "Date of Study" "Date of birth"
## [7] "Gender" "Disease" "Week"
## [10] "Day" "Time" "Average IOP (mmhg)"
## [13] "Hyperemia Level" "Corneal Thickness" "Pupil Size"table(dts$Dog,dts$Treatment)##
## 1 2
## Kepler 60 60
## Mabel 60 60
## Mokie 60 60
## Napoleon 60 60
## Nefertiti 60 60
## Newton 60 60
## Nightingale 60 60
## Oasis 60 60
## Oliver 60 60
## Orion 60 60
## Otto 60 60table(dts$Dog,dts$Week)##
## 1 2 3 4
## Kepler 30 30 30 30
## Mabel 30 30 30 30
## Mokie 30 30 30 30
## Napoleon 30 30 30 30
## Nefertiti 30 30 30 30
## Newton 30 30 30 30
## Nightingale 30 30 30 30
## Oasis 30 30 30 30
## Oliver 30 30 30 30
## Orion 30 30 30 30
## Otto 30 30 30 30ggplot(dts,aes(x=Week,y=`Corneal Thickness`,
color=as.character(Treatment)))+geom_point()+geom_smooth()## `geom_smooth()` using method = 'loess' and formula 'y ~ x'
ggplot(dts,aes(x=Week,y=`Average IOP (mmhg)`,
color=as.character(Treatment)))+geom_point()+geom_smooth()## `geom_smooth()` using method = 'loess' and formula 'y ~ x'
ggplot(dts,aes(x=Week,y=`Average IOP (mmhg)`,
color=as.character(Treatment)))+geom_point()+geom_smooth() + facet_grid(rows = vars(`Dog`),cols=vars(`Time`))## `geom_smooth()` using method = 'loess' and formula 'y ~ x'
Pairwise comparison between Treatments at each timepoint (day*time) Linear mixed model that accounts for Dog (random) eye (fixed)
## [1] "`Average IOP (mmhg)` ~ Treatment+Dog+Eye"
## [1] "`Hyperemia Level` ~ Treatment+Dog+Eye"## Warning in summary.lm(x): essentially perfect fit: summary may be unreliable
## Warning in summary.lm(x): essentially perfect fit: summary may be unreliable
## Warning in summary.lm(x): essentially perfect fit: summary may be unreliable
## Warning in summary.lm(x): essentially perfect fit: summary may be unreliable
## Warning in summary.lm(x): essentially perfect fit: summary may be unreliable
## Warning in summary.lm(x): essentially perfect fit: summary may be unreliable## [1] "`Pupil Size` ~ Treatment+Dog+Eye"## [1] "results are in an excel file"Comparison of average per week
Averae response per week (over days and time of day) for each dog & treatment. Linear mixed model to compare everything to week 1.
## # A tibble: 1,320 x 13
## # Groups: Week, Dog, Time [132]
## Dog `Age (years)` Eye Treatment Gender Disease Week Day Time
## <chr> <dbl> <chr> <fct> <chr> <chr> <dbl> <dbl> <dbl>
## 1 Kepler 3 OD 2 F mutant 1 1 7
## 2 Kepler 3 OD 2 F mutant 1 1 11
## 3 Kepler 3 OD 2 F mutant 1 1 3
## 4 Kepler 3 OD 2 F mutant 1 2 7
## 5 Kepler 3 OD 2 F mutant 1 2 11
## 6 Kepler 3 OD 2 F mutant 1 2 3
## 7 Kepler 3 OD 2 F mutant 1 3 7
## 8 Kepler 3 OD 2 F mutant 1 3 11
## 9 Kepler 3 OD 2 F mutant 1 3 3
## 10 Kepler 3 OD 2 F mutant 1 4 7
## # ... with 1,310 more rows, and 4 more variables: Average IOP (mmhg) <dbl>,
## # Hyperemia Level <dbl>, Corneal Thickness <dbl>, Pupil Size <dbl>## `summarise()` has grouped output by 'Week', 'Dog'. You can override using the `.groups` argument.## [1] "IOP:"## Time term estimate std.error statistic p.value
## 12 3 Week2 0.04787879 0.4613162 0.1037874 0.918028839
## 13 3 Week3 0.61238095 0.4613162 1.3274648 0.194368815
## 14 3 Week4 1.66922078 0.4613162 3.6183878 0.001076804
## 121 7 Week2 -1.57751515 0.7913745 -1.9933864 0.055381592
## 131 7 Week3 -0.75161732 0.7913745 -0.9497619 0.349822359
## 141 7 Week4 1.65824242 0.7913745 2.0953953 0.044679168
## 122 11 Week2 -0.98909091 0.5822760 -1.6986635 0.099732360
## 132 11 Week3 0.47242424 0.5822760 0.8113408 0.423560680
## 142 11 Week4 1.56645022 0.5822760 2.6902196 0.011553093## [1] "Pupil Size:"## Time term estimate std.error statistic p.value
## 12 3 Week2 -0.45454545 0.1102714 -4.1220615 0.0002727408
## 13 3 Week3 0.10909091 0.1102714 0.9892948 0.3304314472
## 14 3 Week4 0.05454545 0.1102714 0.4946474 0.6244522221
## 121 7 Week2 -0.55454545 0.1520294 -3.6476191 0.0009955484
## 131 7 Week3 -0.53636364 0.1520294 -3.5280250 0.0013707889
## 141 7 Week4 -0.04545455 0.1520294 -0.2989852 0.7670127909
## 122 11 Week2 -0.49090909 0.1163849 -4.2179777 0.0002090919
## 132 11 Week3 0.01818182 0.1163849 0.1562214 0.8769050451
## 142 11 Week4 -0.10000000 0.1163849 -0.8592177 0.3970342972## [1] "Hyperemia:"## Time term estimate std.error statistic p.value
## 12 3 Week2 0.4545455 0.10697442 4.249104 1.917719e-04
## 13 3 Week3 1.0727273 0.10697442 10.027886 4.287768e-11
## 14 3 Week4 0.6363636 0.10697442 5.948746 1.609183e-06
## 121 7 Week2 0.2545455 0.09648078 2.638302 1.308152e-02
## 131 7 Week3 0.6818182 0.09648078 7.066881 7.402611e-08
## 141 7 Week4 0.5818182 0.09648078 6.030405 1.280711e-06
## 122 11 Week2 0.4181818 0.12201251 3.427368 1.789790e-03
## 132 11 Week3 0.8636364 0.12201251 7.078261 7.178656e-08
## 142 11 Week4 0.5272727 0.12201251 4.321465 1.567892e-04## [1] "Corneal Thickness:"## Estimate Std. Error t value Pr(>|t|)
## Week2 -3.0454545 4.362263 -0.6981364 0.4904683
## Week3 -0.5909091 4.362263 -0.1354593 0.8931536
## Week4 -11.2727273 4.362263 -2.5841466 0.0148744