Last updated: 2021-06-19

Checks: 7 0

Knit directory: ~/Box/RProjects/pair_con_select/

This reproducible R Markdown analysis was created with workflowr (version 1.6.2). The Checks tab describes the reproducibility checks that were applied when the results were created. The Past versions tab lists the development history.

R Markdown file: up-to-date

Great! Since the R Markdown file has been committed to the Git repository, you know the exact version of the code that produced these results.

Environment: empty

Great job! The global environment was empty. Objects defined in the global environment can affect the analysis in your R Markdown file in unknown ways. For reproduciblity it’s best to always run the code in an empty environment.

Seed: set.seed(20190211)

The command set.seed(20190211) was run prior to running the code in the R Markdown file. Setting a seed ensures that any results that rely on randomness, e.g. subsampling or permutations, are reproducible.

Session information: recorded

Great job! Recording the operating system, R version, and package versions is critical for reproducibility.

Cache: none

Nice! There were no cached chunks for this analysis, so you can be confident that you successfully produced the results during this run.

File paths: relative

Great job! Using relative paths to the files within your workflowr project makes it easier to run your code on other machines.

Repository version: 3155461

Great! You are using Git for version control. Tracking code development and connecting the code version to the results is critical for reproducibility.

The results in this page were generated with repository version 3155461. See the Past versions tab to see a history of the changes made to the R Markdown and HTML files.

Note that you need to be careful to ensure that all relevant files for the analysis have been committed to Git prior to generating the results (you can use wflow_publish or wflow_git_commit). workflowr only checks the R Markdown file, but you know if there are other scripts or data files that it depends on. Below is the status of the Git repository when the results were generated: 


Ignored files:
    Ignored:    .Rhistory
    Ignored:    .Rproj.user/
    Ignored:    analysis/.RData
    Ignored:    analysis/.Rhistory
    Ignored:    analysis/.Rproj.user/
    Ignored:    data/depmap_alkati/Data_Raw/CCLE/CCLE_RNAseq_ExonUsageRatio_20180929.gct
    Ignored:    data/skmel28_sos1_mekq56p_vemurafenib.csv.sb-ea24b981-dvFz4V/
    Ignored:    data/tcga_brca_expression/
    Ignored:    data/tcga_luad_expression/
    Ignored:    data/tcga_skcm_expression/
    Ignored:    output/alkati_filtercutoff_allfilters.csv

Untracked files:
    Untracked:  paircon_boxplot_alkati.pdf

Note that any generated files, e.g. HTML, png, CSS, etc., are not included in this status report because it is ok for generated content to have uncommitted changes.

These are the previous versions of the repository in which changes were made to the R Markdown (analysis/updated_resampling_strategy2.Rmd) and HTML (docs/updated_resampling_strategy2.html) files. If you’ve configured a remote Git repository (see ?wflow_git_remote), click on the hyperlinks in the table below to view the files as they were in that past version.

File Version Author Date Message
Rmd 3155461 haiderinam 2021-06-19 updated method for pairwise comparisons
Rmd 8520286 haiderinam 2021-06-14 Updated website contents on 061421
Rmd 58c60f1 haiderinam 2021-06-10 061021 Updates 
# rm(list=ls())
library(knitr)
library(tictoc)
library(workflowr)
library(VennDiagram)
Loading required package: grid
Loading required package: futile.logger
library(dplyr)
Warning: package 'dplyr' was built under R version 4.0.2

Attaching package: 'dplyr'
The following objects are masked from 'package:stats':

    filter, lag
The following objects are masked from 'package:base':

    intersect, setdiff, setequal, union
library(foreach)
library(doParallel)
Loading required package: iterators
Loading required package: parallel
library(ggplot2)
Warning: package 'ggplot2' was built under R version 4.0.2
library(reshape2)
library(RColorBrewer)
library(devtools)
Loading required package: usethis
Error in get(genname, envir = envir) : object 'testthat_print' not found
library(ggsignif)
library(plotly)

Attaching package: 'plotly'
The following object is masked from 'package:ggplot2':

    last_plot
The following object is masked from 'package:stats':

    filter
The following object is masked from 'package:graphics':

    layout
library(BiocManager)
Bioconductor version 3.12 (BiocManager 1.30.10), ?BiocManager::install for help
Bioconductor version '3.12' is out-of-date; the current release version '3.13'
  is available with R version '4.1'; see https://bioconductor.org/install

Attaching package: 'BiocManager'
The following object is masked from 'package:devtools':

    install
# library(ensembldb) #Loading this with Dplyr commands seems to throw an error in Rmd
# library(EnsDb.Hsapiens.v86) #Loading this with Dplyr commands seems to throw an error in Rmd

source("code/contab_maker.R")
source("code/alldata_compiler.R")
source("code/quadratic_solver.R")
source("code/mut_excl_genes_generator.R")
source("code/mut_excl_genes_datapoints.R")
source("code/simresults_generator.R")

# source("../code/contab_maker.R")
# source("../code/alldata_compiler.R")
# source("../code/quadratic_solver.R")
# source("../code/mut_excl_genes_generator.R")
# source("../code/mut_excl_genes_datapoints.R")
# source("../code/simresults_generator.R")


######################Cleanup for GGPlot2#########################################
cleanup=theme_bw() +
  theme(plot.title = element_text(hjust=.5),
        panel.grid.major = element_blank(),
        panel.grid.major.y = element_blank(),
        panel.background = element_blank(),
        # axis.line = element_line(color = "black"),
        axis.text = element_text(face="bold",color="black",size="11"),
        text=element_text(size=11,face="bold"),
        axis.title=element_text(face="bold",size="11"))
  nameposctrl1<-'BRAF'
  #Positive control 1
  nameposctrl2<-'NRAS'
  #Oncogene in Question
  namegene<-'ATI'
  #Mutation Boolean (Y or N)
  mtn<-'N'
  #Name Mutation for Positive Ctrl 1
  nameposctrl1mt<-'V600E'
  #Name of Mutation for Positive Ctrl 2
  nameposctrl2mt<-'Q61L'

alldata=read.csv("output/all_data_skcm.csv",sep=",",header=T,stringsAsFactors=F)
# alldata=read.csv("../output/all_data_skcm.csv",sep=",",header=T,stringsAsFactors=F)
head(alldata)
  X        Patid mean_RPKM_1.19 mean_RPKM_20.29 Ratio20.29 mRNA_count  BRAF
1 1 TCGA-BF-A1PU     0.62445977      1.24042009   1.986389        948 V600E
2 2 TCGA-BF-A1PV     0.02099345      0.15815619   7.533598         82   NaN
3 3 TCGA-BF-A1PX     0.01752838      0.09612414   5.483914         92 V600E
4 4 TCGA-BF-A1PZ     0.19874434      7.27553619  36.607514       2822   NaN
5 5 TCGA-BF-A1Q0     2.13353636      3.71661801   1.741999       2211   NaN
6 6 TCGA-BF-A3DJ     0.06244694      0.55656239   8.912565        281 V600E
  NRAS RSEM_normalized ATI
1  NaN        107.1429   0
2 Q61L          8.9659   0
3  NaN         14.5985   0
4 Q61R        329.0810   1
5  NaN        277.0434   0
6  NaN         35.1542   0

Contab downsampler function that the next code chunk uses:

contab_downsampler<- function(contab,prob) {
  p11=contab[1,1]
  p10=contab[1,2]
  p01=contab[2,1]
  p00=contab[2,2]
  contab_downsampled=contab ###Just making a dummy contingency table
  contab_downsampled[1,1]=prob*p11/(p10+p11)
  contab_downsampled[1,2]=prob*p10/(p10+p11)
  contab_downsampled[2,1]=(1-prob)*p01/(p00+p01)
  contab_downsampled[2,2]=(1-prob)*p00/(p00+p01)
  contab_downsampled
}
mut_excl_genes_generator2=function(cohort_size,incidence,or_pair1,or_pair2){
  # or_pair1=.9
  # or_pair2=.01
  # 
  # incidence=40
  # cohort_size=1000
  
  or_p=or_pair2
  or_q=or_pair1
  q00=0.05
  p00=0.05
  ###PC1 vs GOI pair###
  p_goi=incidence/cohort_size
  
  q10=(1-q00)-p_goi
  q11=(p_goi*(1/q00)*or_q*((1-q00)-p_goi))/(1+(1/q00)*or_q*((1-q00)-p_goi))
  q01=(p_goi)/(1+(1/q00)*or_q*((1-q00)-p_goi))
  
  ###PC1 vs PC2 pair###
  p_pc1=q11+q10
  
  p01=(1-p00)-p_pc1
  p11=(p_pc1*(1/p00)*or_p*((1-p00)-p_pc1))/(1+((1/p00)*or_p*((1-p00)-p_pc1)))
  p10=(p_pc1)/(1+(1/p00)*or_p*((1-p00)-p_pc1))

  # gene_pair_1=(as.numeric(c(q11,q10,q01,q00)))
  # gene_pair_2=(as.numeric(c(p11,p10,p01,p00)))
  gene_pair_1=(as.numeric(c(q11,q01,q10,q00)))
  gene_pair_2=(as.numeric(c(p11,p10,p01,p00)))
  return(list(gene_pair_1,gene_pair_2))
}

# mut_excl_genes_generator2(500,12,.1,.05)
# mut_excl_genes_generator2(500,36,.1,.05)
# mut_excl_genes_generator(500,12,.1,.05)
# mut_excl_genes_generator(500,36,.1,.05)

Another function that takes in a contingency table, and resamples is based on its probabilities on a multinomial distribution. Number of resampling simulations are an input of the function

contab_simulator=function(contab,nsims,cohort_size){
  contabs_resampled=rmultinom(nsims, #number of datasets to generate
          cohort_size, #sample size of the total population
          c(contab[1,1],
            contab[1,2],
            contab[2,1],
            contab[2,2])) #probability of positive event
  contabs_resampled_t=t(contabs_resampled)
  colnames(contabs_resampled_t)=c("p11","p10","p01","p00")
  contabs_resampled_t
}

# contab_simulator=function(contab,nsims,cohort_size){
#   contabs_resampled=rmultinom(nsims, #number of datasets to generate
#           1, #sample size of the total population
#           c(contab[1,1],
#             contab[1,2],
#             contab[2,1],
#             contab[2,2])) #probability of positive event
#   contabs_resampled_t=t(contabs_resampled)
#   colnames(contabs_resampled_t)=c("p11","p10","p01","p00")
#   contabs_resampled_t
# }
# contab=pc1pc2_contab_probabilities
# nsims=1000
# cohort_size=cohort_size_curr
# pc1rawpc2_contabs_sims=contab_simulator(pc1pc2_contab_probabilities,1000,cohort_size_curr)
# rm(list=ls())
###Not mutation specific generation of counts###
alldata_comp=alldata_compiler(alldata,nameposctrl1,nameposctrl2,namegene,'N',"N/A","N/A")[[2]]
head(alldata_comp)
  X        Patid Positive_Ctrl1 Positive_Ctrl2 genex rndmarray
1 1 TCGA-BF-A1PU              1              0     0         0
2 2 TCGA-BF-A1PV              0              1     0         1
3 3 TCGA-BF-A1PX              1              0     0         0
4 4 TCGA-BF-A1PZ              0              1     1         0
5 5 TCGA-BF-A1Q0              0              0     0         0
6 6 TCGA-BF-A3DJ              1              0     0         0
###Calculating Odds ratios and GOI frequencies for the raw data###
cohort_size=length(alldata_comp$Positive_Ctrl1)
pc1pc2_contab_counts=contab_maker(alldata_comp$Positive_Ctrl1,alldata_comp$Positive_Ctrl2,alldata_comp)[2:1, 2:1]
goipc1_contab_counts=contab_maker(alldata_comp$genex,alldata_comp$Positive_Ctrl1,alldata_comp)[2:1, 2:1]
###Had to add the 2:1 bits because the contab maker spits out NN YY whereas we wanted YNYN
goipc2_contab_counts=contab_maker(alldata_comp$genex,alldata_comp$Positive_Ctrl2,alldata_comp)[2:1, 2:1]

# pc1pc2_contab_counts=gene_pair_2_table
# goipc1_contab_counts=contab_maker(alldata_comp$genex,alldata_comp$Positive_Ctrl1,alldata_comp)[2:1, 2:1]
# goipc1_contab_counts=gene_pair_1_table

cohort_size_curr=cohort_size
# goipc2_contab_counts=contab_maker(alldata_comp$genex,alldata_comp$Positive_Ctrl2,alldata_comp)[2:1, 2:1]
pc1pc2_contab_probabilities=pc1pc2_contab_counts/cohort_size_curr
goipc1_contab_probabilities=goipc1_contab_counts/cohort_size_curr
goipc2_contab_probabilities=goipc2_contab_counts/cohort_size_curr
# pc1pc2_contab_probabilities=pc1pc2_contab_counts
# goipc1_contab_probabilities=goipc1_contab_counts
# goipc2_contab_probabilities=goipc2_contab_counts/cohort_size

or_pc1pc2=pc1pc2_contab_probabilities[1,1]*pc1pc2_contab_probabilities[2,2]/(pc1pc2_contab_probabilities[1,2]*pc1pc2_contab_probabilities[2,1])
or_goipc1=goipc1_contab_probabilities[1,1]*goipc1_contab_probabilities[2,2]/(goipc1_contab_probabilities[1,2]*goipc1_contab_probabilities[2,1])
or_goipc2=goipc2_contab_probabilities[1,1]*goipc2_contab_probabilities[2,2]/(goipc2_contab_probabilities[1,2]*goipc2_contab_probabilities[2,1])

goi_freq=goipc1_contab_probabilities[1,1]+goipc1_contab_probabilities[1,2]
goi_freq=.25
# class(goi_freq)

###

###Downsampling PC1 to the probability of GOI without changing ORs###
###The function below converts contingency table data to a new contingency table in which the data is downsampled to the desired frequency, aka the frequency of the GOI in this case###
pc1new_pc2_contab=contab_downsampler(pc1pc2_contab_probabilities,goi_freq)
goinew_pc1_contab=contab_downsampler(goipc1_contab_probabilities,goi_freq)
goinew_pc2_contab=contab_downsampler(goipc2_contab_probabilities,goi_freq)
##original contab:
head(pc1pc2_contab_probabilities)
           [,1]      [,2]
[1,] 0.01424501 0.4985755
[2,] 0.25925926 0.2279202
###downsampled contab:
head(pc1new_pc2_contab)
            [,1]      [,2]
[1,] 0.006944444 0.2430556
[2,] 0.399122807 0.3508772
pc1rawpc2_contabs_sims=contab_simulator(pc1pc2_contab_probabilities,1000,cohort_size_curr)
pc1pc2_contabs_sims=contab_simulator(pc1new_pc2_contab,1000,cohort_size_curr)
goipc1_contabs_sims=contab_simulator(goinew_pc1_contab,1000,cohort_size_curr)
goipc2_contabs_sims=contab_simulator(goinew_pc2_contab,1000,cohort_size_curr)
# goipc2_contabs_sims=contab_simulator(goinew_pc2_contab,1000,cohort_size)
# head(pc1pc2_contabs_sims) #each row in this dataset is a new contab
pc1rawpc2_contabs_sims=data.frame(pc1rawpc2_contabs_sims)
  pc1rawpc2_contabs_sims=pc1rawpc2_contabs_sims%>%
  mutate(or=p11*p00/(p10*p01))

pc1pc2_contabs_sims=data.frame(pc1pc2_contabs_sims)
  pc1pc2_contabs_sims=pc1pc2_contabs_sims%>%
  mutate(or=p11*p00/(p10*p01))
  
goipc1_contabs_sims=data.frame(goipc1_contabs_sims)
  goipc1_contabs_sims=goipc1_contabs_sims%>%
  mutate(or=p11*p00/(p10*p01))
  
goipc2_contabs_sims=data.frame(goipc2_contabs_sims)
goipc2_contabs_sims=goipc2_contabs_sims%>%
mutate(or=p11*p00/(p10*p01))
  
# goipc2_contabs_sims=data.frame(goipc2_contabs_sims)
#   goipc2_contabs_sims=goipc2_contabs_sims%>%
#   mutate(or=p11*p00/(p10*p01))
pc1rawpc2_contabs_sims$comparison="pc1rawpc2"
pc1pc2_contabs_sims$comparison="pc1pc2"
goipc1_contabs_sims$comparison="goipc1"
goipc2_contabs_sims$comparison="goipc2"
or_median_raw=quantile(pc1rawpc2_contabs_sims$or,na.rm = T)[3]
or_uq_raw=quantile(pc1rawpc2_contabs_sims$or,na.rm = T)[4]
or_median_downsampled=quantile(pc1pc2_contabs_sims$or,na.rm = T)[3]
or_uq_downsampled=quantile(pc1pc2_contabs_sims$or,na.rm = T)[4]

pc1rawpc2_contabs_sims=pc1rawpc2_contabs_sims%>%
  mutate(isgreater_raw_median=case_when(or>or_median_raw~1,
                             TRUE~0),
         isgreater_raw_uq=case_when(or>or_uq_raw~1,
                             TRUE~0),
         isgreater_median=case_when(or>or_median_downsampled~1,
                             TRUE~0),
         isgreater_uq=case_when(or>or_uq_downsampled~1,
                             TRUE~0)
         )
pc1pc2_contabs_sims=pc1pc2_contabs_sims%>%
  mutate(isgreater_raw_median=case_when(or>or_median_raw~1,
                             TRUE~0),
         isgreater_raw_uq=case_when(or>or_uq_raw~1,
                             TRUE~0),
         isgreater_median=case_when(or>or_median_downsampled~1,
                             TRUE~0),
         isgreater_uq=case_when(or>or_uq_downsampled~1,
                             TRUE~0)
         )
goipc1_contabs_sims=goipc1_contabs_sims%>%
  mutate(isgreater_raw_median=case_when(or>or_median_raw~1,
                             TRUE~0),
         isgreater_raw_uq=case_when(or>or_uq_raw~1,
                             TRUE~0),
         isgreater_median=case_when(or>or_median_downsampled~1,
                             TRUE~0),
         isgreater_uq=case_when(or>or_uq_downsampled~1,
                             TRUE~0)
         )

goipc2_contabs_sims=goipc2_contabs_sims%>%
  mutate(isgreater_raw_median=case_when(or>or_median_raw~1,
                             TRUE~0),
         isgreater_raw_uq=case_when(or>or_uq_raw~1,
                             TRUE~0),
         isgreater_median=case_when(or>or_median_downsampled~1,
                             TRUE~0),
         isgreater_uq=case_when(or>or_uq_downsampled~1,
                             TRUE~0)
         )
# pc1pc2_contabs_sims=pc1pc2_contabs_sims%>%
#   mutate(isgreater=case_when(or>=or_pc1pc2~1,
#                              TRUE~0))
# goipc1_contabs_sims=goipc1_contabs_sims%>%
#   mutate(isgreater=case_when(or>=or_pc1pc2~1,
#                              TRUE~0))
# goipc2_contabs_sims=goipc2_contabs_sims%>%
#   mutate(isgreater=case_when(or>=or_pc1pc2~1,
#                              TRUE~0))
pc1rawpc2_isgreater_raw_median=sum(pc1rawpc2_contabs_sims$isgreater_raw_median)
pc1rawpc2_isgreater_raw_uq=sum(pc1rawpc2_contabs_sims$isgreater_raw_uq)
pc1rawpc2_isgreater_median=sum(pc1rawpc2_contabs_sims$isgreater_median)
pc1rawpc2_isgreater_uq=sum(pc1rawpc2_contabs_sims$isgreater_uq)

pc1pc2_isgreater_raw_median=sum(pc1pc2_contabs_sims$isgreater_raw_median)
pc1pc2_isgreater_raw_uq=sum(pc1pc2_contabs_sims$isgreater_raw_uq)
pc1pc2_isgreater_median=sum(pc1pc2_contabs_sims$isgreater_median)
pc1pc2_isgreater_uq=sum(pc1pc2_contabs_sims$isgreater_uq)

goipc1_isgreater_raw_median=sum(goipc1_contabs_sims$isgreater_raw_median)
goipc1_isgreater_raw_uq=sum(goipc1_contabs_sims$isgreater_raw_uq)
goipc1_isgreater_median=sum(goipc1_contabs_sims$isgreater_median)
goipc1_isgreater_uq=sum(goipc1_contabs_sims$isgreater_uq)


plotting_df=rbind(pc1pc2_contabs_sims,goipc1_contabs_sims,goipc2_contabs_sims)
# plotting_df=rbind(pc1pc2_contabs_sims,goipc1_contabs_sims)
# 


ggplot(plotting_df,aes(x=(or),fill=comparison))+
  geom_histogram(bins=40,alpha=0.55,position="identity")+
  # geom_histogram(bins=50,alpha=0.55)+
  scale_y_continuous(expand=c(0,0),name="Count")+
  scale_x_continuous(expand=c(0,0),trans="log10",name="Odds Ratio")+
  scale_fill_brewer(palette="Set2")+
  # geom_vline(xintercept = or_pc1pc2)+
  cleanup
Warning: Transformation introduced infinite values in continuous x-axis
Warning: Removed 86 rows containing non-finite values (stat_bin).

ggplot(plotting_df,aes(y=(or),x=comparison),fill=factor(comparison))+
  geom_boxplot()+
  scale_y_continuous(name="Odds Ratio",trans="log10")+
  scale_x_discrete(name="")+
  scale_fill_brewer(palette="Set2")+
  geom_hline(yintercept = or_uq_downsampled,linetype="dashed")+
  cleanup+
  theme(legend.position = "none",
        axis.ticks.x = element_blank())
Warning: Transformation introduced infinite values in continuous y-axis
Warning: Removed 86 rows containing non-finite values (stat_boxplot).

# ggsave("paircon_boxplot.pdf",width = 3,height=2,units="in",useDingbats=F)

ggplot(plotting_df%>%filter(comparison%in%"pc1pc2"),aes(x=or,fill=comparison))+
  geom_histogram(bins=50,alpha=0.55,position="identity")+
  geom_density(alpha=.6)+
  scale_y_continuous(expand=c(0,0),name="Count")+
  scale_x_continuous(expand=c(0,0),name="Odds Ratio")+
  # geom_vline(xintercept = or_pc1pc2)+
  cleanup

# rm(list=ls())
###Not mutation specific generation of counts###
alldata_comp=alldata_compiler(alldata,nameposctrl1,nameposctrl2,namegene,'N',"N/A","N/A")[[2]]

###Mutation specific
####Decided to not look at mutation specific counts too closely because co-occurrence of brafv600e and nrasq61L is a VERY high bar
# alldata_comp=alldata_compiler(alldata,nameposctrl1,nameposctrl2,namegene,'Y',"V600E","Q61L")[[2]]
head(alldata_comp)
  X        Patid Positive_Ctrl1 Positive_Ctrl2 genex rndmarray
1 1 TCGA-BF-A1PU              1              0     0         0
2 2 TCGA-BF-A1PV              0              1     0         1
3 3 TCGA-BF-A1PX              1              0     0         0
4 4 TCGA-BF-A1PZ              0              1     1         0
5 5 TCGA-BF-A1Q0              0              0     0         0
6 6 TCGA-BF-A3DJ              1              0     0         0
###Calculating Odds ratios and GOI frequencies for the raw data###
cohort_size=length(alldata_comp$Positive_Ctrl1)
pc1pc2_contab_counts=contab_maker(alldata_comp$Positive_Ctrl1,alldata_comp$Positive_Ctrl2,alldata_comp)[2:1, 2:1]
pc1pc2_contab_counts=rbind(c(15,25),c(81,230)) ###NRAS ATI
goipc1_contab_counts=contab_maker(alldata_comp$genex,alldata_comp$Positive_Ctrl1,alldata_comp)[2:1, 2:1]
# goipc1_contab_counts=rbind(c(1,110),c(186,365)) ###EGFR KRAS
fisher.test(goipc1_contab_counts)

    Fisher's Exact Test for Count Data

data:  goipc1_contab_counts
p-value = 0.4071
alternative hypothesis: true odds ratio is not equal to 1
95 percent confidence interval:
 0.365048 1.535988
sample estimates:
odds ratio 
 0.7531399 
###Had to add the 2:1 bits because the contab maker spits out NN YY whereas we wanted YNYN
goipc2_contab_counts=contab_maker(alldata_comp$genex,alldata_comp$Positive_Ctrl2,alldata_comp)[2:1, 2:1]

# pc1pc2_contab_counts=gene_pair_2_table
# goipc1_contab_counts=contab_maker(alldata_comp$genex,alldata_comp$Positive_Ctrl1,alldata_comp)[2:1, 2:1]
# goipc1_contab_counts=gene_pair_1_table

cohort_size_curr=cohort_size
# goipc2_contab_counts=contab_maker(alldata_comp$genex,alldata_comp$Positive_Ctrl2,alldata_comp)[2:1, 2:1]
pc1pc2_contab_probabilities=pc1pc2_contab_counts/cohort_size_curr
goipc1_contab_probabilities=goipc1_contab_counts/cohort_size_curr
goipc2_contab_probabilities=goipc2_contab_counts/cohort_size_curr
# pc1pc2_contab_probabilities=pc1pc2_contab_counts
# goipc1_contab_probabilities=goipc1_contab_counts
# goipc2_contab_probabilities=goipc2_contab_counts/cohort_size

or_pc1pc2=pc1pc2_contab_probabilities[1,1]*pc1pc2_contab_probabilities[2,2]/(pc1pc2_contab_probabilities[1,2]*pc1pc2_contab_probabilities[2,1])
or_goipc1=goipc1_contab_probabilities[1,1]*goipc1_contab_probabilities[2,2]/(goipc1_contab_probabilities[1,2]*goipc1_contab_probabilities[2,1])
or_goipc2=goipc2_contab_probabilities[1,1]*goipc2_contab_probabilities[2,2]/(goipc2_contab_probabilities[1,2]*goipc2_contab_probabilities[2,1])

goi_freq=goipc1_contab_probabilities[1,1]+goipc1_contab_probabilities[1,2]
# goi_freq=.25
# class(goi_freq)

###

###Downsampling PC1 to the probability of GOI without changing ORs###
###The function below converts contingency table data to a new contingency table in which the data is downsampled to the desired frequency, aka the frequency of the GOI in this case###
pc1new_pc2_contab=contab_downsampler(pc1pc2_contab_probabilities,goi_freq)
goinew_pc1_contab=contab_downsampler(goipc1_contab_probabilities,goi_freq)
goinew_pc2_contab=contab_downsampler(goipc2_contab_probabilities,goi_freq)
##original contab:
head(pc1pc2_contab_probabilities)
           [,1]       [,2]
[1,] 0.04273504 0.07122507
[2,] 0.23076923 0.65527066
###downsampled contab:
head(pc1new_pc2_contab)
           [,1]       [,2]
[1,] 0.04273504 0.07122507
[2,] 0.23076923 0.65527066
pc1rawpc2_contabs_sims=contab_simulator(pc1pc2_contab_probabilities,1000,cohort_size_curr)
pc1pc2_contabs_sims=contab_simulator(pc1new_pc2_contab,1000,cohort_size_curr)
goipc1_contabs_sims=contab_simulator(goinew_pc1_contab,1000,cohort_size_curr)
goipc2_contabs_sims=contab_simulator(goinew_pc2_contab,1000,cohort_size_curr)
# goipc2_contabs_sims=contab_simulator(goinew_pc2_contab,1000,cohort_size)
# head(pc1pc2_contabs_sims) #each row in this dataset is a new contab
pc1rawpc2_contabs_sims=data.frame(pc1rawpc2_contabs_sims)
  pc1rawpc2_contabs_sims=pc1rawpc2_contabs_sims%>%
  mutate(or=p11*p00/(p10*p01))

pc1pc2_contabs_sims=data.frame(pc1pc2_contabs_sims)
  pc1pc2_contabs_sims=pc1pc2_contabs_sims%>%
  mutate(or=p11*p00/(p10*p01))
  
goipc1_contabs_sims=data.frame(goipc1_contabs_sims)
  goipc1_contabs_sims=goipc1_contabs_sims%>%
  mutate(or=p11*p00/(p10*p01))
  
goipc2_contabs_sims=data.frame(goipc2_contabs_sims)
goipc2_contabs_sims=goipc2_contabs_sims%>%
mutate(or=p11*p00/(p10*p01))
  
# goipc2_contabs_sims=data.frame(goipc2_contabs_sims)
#   goipc2_contabs_sims=goipc2_contabs_sims%>%
#   mutate(or=p11*p00/(p10*p01))
pc1rawpc2_contabs_sims$comparison="pc1rawpc2"
pc1pc2_contabs_sims$comparison="pc1pc2"
goipc1_contabs_sims$comparison="goipc1"
goipc2_contabs_sims$comparison="goipc2"
or_median_raw=quantile(pc1rawpc2_contabs_sims$or,na.rm = T)[3]
or_uq_raw=quantile(pc1rawpc2_contabs_sims$or,na.rm = T)[4]
or_median_downsampled=quantile(pc1pc2_contabs_sims$or,na.rm = T)[3]
or_uq_downsampled=quantile(pc1pc2_contabs_sims$or,na.rm = T)[4]

pc1rawpc2_contabs_sims=pc1rawpc2_contabs_sims%>%
  mutate(isgreater_raw_median=case_when(or>=or_median_raw~1,
                             TRUE~0),
         isgreater_raw_uq=case_when(or>or_uq_raw~1,
                             TRUE~0),
         isgreater_median=case_when(or>or_median_downsampled~1,
                             TRUE~0),
         isgreater_uq=case_when(or>or_uq_downsampled~1,
                             TRUE~0)
         )
pc1pc2_contabs_sims=pc1pc2_contabs_sims%>%
  mutate(isgreater_raw_median=case_when(or>or_median_raw~1,
                             TRUE~0),
         isgreater_raw_uq=case_when(or>or_uq_raw~1,
                             TRUE~0),
         isgreater_median=case_when(or>or_median_downsampled~1,
                             TRUE~0),
         isgreater_uq=case_when(or>or_uq_downsampled~1,
                             TRUE~0)
         )
goipc1_contabs_sims=goipc1_contabs_sims%>%
  mutate(isgreater_raw_median=case_when(or>or_median_raw~1,
                             TRUE~0),
         isgreater_raw_uq=case_when(or>or_uq_raw~1,
                             TRUE~0),
         isgreater_median=case_when(or>or_median_downsampled~1,
                             TRUE~0),
         isgreater_uq=case_when(or>or_uq_downsampled~1,
                             TRUE~0)
         )

goipc2_contabs_sims=goipc2_contabs_sims%>%
  mutate(isgreater_raw_median=case_when(or>or_median_raw~1,
                             TRUE~0),
         isgreater_raw_uq=case_when(or>or_uq_raw~1,
                             TRUE~0),
         isgreater_median=case_when(or>or_median_downsampled~1,
                             TRUE~0),
         isgreater_uq=case_when(or>or_uq_downsampled~1,
                             TRUE~0)
         )
# pc1pc2_contabs_sims=pc1pc2_contabs_sims%>%
#   mutate(isgreater=case_when(or>=or_pc1pc2~1,
#                              TRUE~0))
# goipc1_contabs_sims=goipc1_contabs_sims%>%
#   mutate(isgreater=case_when(or>=or_pc1pc2~1,
#                              TRUE~0))
# goipc2_contabs_sims=goipc2_contabs_sims%>%
#   mutate(isgreater=case_when(or>=or_pc1pc2~1,
#                              TRUE~0))
pc1rawpc2_isgreater_raw_median=sum(pc1rawpc2_contabs_sims$isgreater_raw_median)
pc1rawpc2_isgreater_raw_uq=sum(pc1rawpc2_contabs_sims$isgreater_raw_uq)
pc1rawpc2_isgreater_median=sum(pc1rawpc2_contabs_sims$isgreater_median)
pc1rawpc2_isgreater_uq=sum(pc1rawpc2_contabs_sims$isgreater_uq)

pc1pc2_isgreater_raw_median=sum(pc1pc2_contabs_sims$isgreater_raw_median)
pc1pc2_isgreater_raw_uq=sum(pc1pc2_contabs_sims$isgreater_raw_uq)
pc1pc2_isgreater_median=sum(pc1pc2_contabs_sims$isgreater_median)
pc1pc2_isgreater_uq=sum(pc1pc2_contabs_sims$isgreater_uq)

goipc1_isgreater_raw_median=sum(goipc1_contabs_sims$isgreater_raw_median)
goipc1_isgreater_raw_uq=sum(goipc1_contabs_sims$isgreater_raw_uq)
goipc1_isgreater_median=sum(goipc1_contabs_sims$isgreater_median)
goipc1_isgreater_uq=sum(goipc1_contabs_sims$isgreater_uq)


plotting_df=rbind(pc1pc2_contabs_sims,goipc1_contabs_sims,goipc2_contabs_sims)
# plotting_df=rbind(pc1pc2_contabs_sims,goipc1_contabs_sims)
# 

# a=plotting_df%>%filter(comparison%in%"pc1pc2")
ggplot(plotting_df,aes(x=(or),fill=comparison))+
  geom_histogram(bins=40,alpha=0.55,position="identity")+
  # geom_histogram(bins=50,alpha=0.55)+
  scale_y_continuous(expand=c(0,0),name="Count")+
  scale_x_continuous(expand=c(0,0),trans="log10",name="Odds Ratio")+
  scale_fill_brewer(palette="Set2")+
  # geom_vline(xintercept = or_pc1pc2)+
  cleanup

ggplot(plotting_df,aes(y=(or),x=comparison),fill=factor(comparison))+
  geom_boxplot()+
  scale_y_continuous(name="Odds Ratio",trans="log10")+
  scale_x_discrete(name="")+
  scale_fill_brewer(palette="Set2")+
  geom_hline(yintercept = or_uq_downsampled,linetype="dashed")+
  cleanup+
  theme(legend.position = "none",
        axis.ticks.x = element_blank())

ggsave("paircon_boxplot_alkati.pdf",width = 3,height=2,units="in",useDingbats=F)

ggplot(plotting_df%>%filter(comparison%in%"pc1pc2"),aes(x=or,fill=comparison))+
  geom_histogram(bins=50,alpha=0.55,position="identity")+
  geom_density(alpha=.6)+
  scale_y_continuous(expand=c(0,0),name="Count")+
  scale_x_continuous(expand=c(0,0),name="Odds Ratio")+
  # geom_vline(xintercept = or_pc1pc2)+
  cleanup

# rm(list=ls())
or_pair1=c(.01,.05,.1,.2,.3,.4,.5,.6,.7,.8,.9,1)
or_pair2=seq(.01,.2,by=.02)
incidence=seq(4,36,by=2)
cohort_size=seq(100,1000,by=200)

# or_pair1=c(.01,.05)
# or_pair2=seq(.01,.1,by=.05)
# incidence=seq(4,36,by=20)
# cohort_size=seq(100,1000,by=400)

    # or_pair1=.05
    # or_pair2=.01
    # incidence=12
    # cohort_size=500
    # i=1
    # j=1
    # k=1
    # l=1

true_or_vals=or_pair1
cohort_size_vals=cohort_size
gene1_total_vals=incidence

simresults_compiled_alldata=as.list(length(or_pair2)*length(or_pair1)*length(incidence)*length(cohort_size))
simresults_compiled=matrix(,length(or_pair2)*length(or_pair1)*length(incidence)*length(cohort_size),ncol=16)
ct=1
for(j in 1:length(incidence)){
  tic()
for(l in 1:length(or_pair2)){
  for(k in 1:length(cohort_size_vals)){
    for(i in 1:length(true_or_vals)){
gene_pair_1=unlist(mut_excl_genes_generator(cohort_size[k],incidence[j],or_pair1[i],or_pair2[l])[1])
gene_pair_1_table=rbind(c(gene_pair_1[1],gene_pair_1[2]),c(gene_pair_1[3],gene_pair_1[4])) ###make sure these are the right indices
gene_pair_2=unlist(mut_excl_genes_generator(cohort_size[k],incidence[j],or_pair1[i],or_pair2[l])[2])
gene_pair_2_table=rbind(c(gene_pair_2[1],gene_pair_2[2]),c(gene_pair_2[3],gene_pair_2[4]))
# alldata_1=mut_excl_genes_datapoints(gene_pair_1)
# 
# alldata_2=mut_excl_genes_datapoints(gene_pair_2)
# alldata_comp_1=alldata_compiler(alldata_1,"gene2","gene3","gene1",'N',"N/A","N/A")[[2]]
# 
# genex_replication_prop_1=alldata_compiler(alldata_1,"gene2","gene3","gene1",'N',"N/A","N/A")[[1]]
# alldata_comp_2=alldata_compiler(alldata_2,"gene2","gene3","gene1",'N',"N/A","N/A")[[2]]
# genex_replication_prop_2=alldata_compiler(alldata_2,"gene2","gene3","gene1",'N',"N/A","N/A")[[1]]


###Calculating Odds ratios and GOI frequencies for the raw data###
# cohort_size_curr=length(alldata_comp$Positive_Ctrl1)
cohort_size_curr=cohort_size[k]
# pc1pc2_contab_counts=contab_maker(alldata_comp$Positive_Ctrl1,alldata_comp$Positive_Ctrl2,alldata_comp)[2:1, 2:1]
pc1pc2_contab_counts=gene_pair_2_table
# goipc1_contab_counts=contab_maker(alldata_comp$genex,alldata_comp$Positive_Ctrl1,alldata_comp)[2:1, 2:1]
goipc1_contab_counts=gene_pair_1_table


# goipc2_contab_counts=contab_maker(alldata_comp$genex,alldata_comp$Positive_Ctrl2,alldata_comp)[2:1, 2:1]
# pc1pc2_contab_probabilities=pc1pc2_contab_counts/cohort_size_curr
# goipc1_contab_probabilities=goipc1_contab_counts/cohort_size_curr
pc1pc2_contab_probabilities=pc1pc2_contab_counts
goipc1_contab_probabilities=goipc1_contab_counts
# goipc2_contab_probabilities=goipc2_contab_counts/cohort_size

or_pc1pc2=pc1pc2_contab_probabilities[1,1]*pc1pc2_contab_probabilities[2,2]/(pc1pc2_contab_probabilities[1,2]*pc1pc2_contab_probabilities[2,1])
or_goipc1=goipc1_contab_probabilities[1,1]*goipc1_contab_probabilities[2,2]/(goipc1_contab_probabilities[1,2]*goipc1_contab_probabilities[2,1])
# or_goipc2=goipc2_contab_probabilities[1,1]*goipc2_contab_probabilities[2,2]/(goipc2_contab_probabilities[1,2]*goipc2_contab_probabilities[2,1])

goi_freq=goipc1_contab_probabilities[1,1]+goipc1_contab_probabilities[1,2]
# goi_freq=.01
# class(goi_freq)

###

###Downsampling PC1 to the probability of GOI without changing ORs###
###The function below converts contingency table data to a new contingency table in which the data is downsampled to the desired frequency, aka the frequency of the GOI in this case###
pc1new_pc2_contab=contab_downsampler(pc1pc2_contab_probabilities,goi_freq)
goinew_pc1_contab=contab_downsampler(goipc1_contab_probabilities,goi_freq)
# goinew_pc2_contab=contab_downsampler(goipc2_contab_probabilities,goi_freq)
###original contab:
# head(pc1pc2_contab_probabilities) 
###downsampled contab:
# head(pc1new_pc2_contab) 

pc1rawpc2_contabs_sims=contab_simulator(pc1pc2_contab_probabilities,1000,cohort_size_curr)
pc1pc2_contabs_sims=contab_simulator(pc1new_pc2_contab,1000,cohort_size_curr)
goipc1_contabs_sims=contab_simulator(goinew_pc1_contab,1000,cohort_size_curr)
# goipc2_contabs_sims=contab_simulator(goinew_pc2_contab,1000,cohort_size)
# head(pc1pc2_contabs_sims) #each row in this dataset is a new contab
pc1rawpc2_contabs_sims=data.frame(pc1rawpc2_contabs_sims)
  pc1rawpc2_contabs_sims=pc1rawpc2_contabs_sims%>%
  mutate(or=p11*p00/(p10*p01))

pc1pc2_contabs_sims=data.frame(pc1pc2_contabs_sims)
  pc1pc2_contabs_sims=pc1pc2_contabs_sims%>%
  mutate(or=p11*p00/(p10*p01))
  
goipc1_contabs_sims=data.frame(goipc1_contabs_sims)
  goipc1_contabs_sims=goipc1_contabs_sims%>%
  mutate(or=p11*p00/(p10*p01))
  

  
# goipc2_contabs_sims=data.frame(goipc2_contabs_sims)
#   goipc2_contabs_sims=goipc2_contabs_sims%>%
#   mutate(or=p11*p00/(p10*p01))
pc1rawpc2_contabs_sims$comparison="pc1rawpc2"
pc1pc2_contabs_sims$comparison="pc1pc2"
goipc1_contabs_sims$comparison="goipc1"
# goipc2_contabs_sims$comparison="goipc2"
or_median_raw=quantile(pc1rawpc2_contabs_sims$or,na.rm = T)[3]
or_uq_raw=quantile(pc1rawpc2_contabs_sims$or,na.rm = T)[4]
or_median_downsampled=quantile(pc1pc2_contabs_sims$or,na.rm = T)[3]
or_uq_downsampled=quantile(pc1pc2_contabs_sims$or,na.rm = T)[4]

# pc1rawpc2_contabs_sims=pc1rawpc2_contabs_sims%>%
#   mutate(isgreater_raw_median=case_when(or>=or_median_raw~1,
#                              TRUE~0),
#          isgreater_raw_uq=case_when(or>=or_uq_raw~1,
#                              TRUE~0),
#          isgreater_median=case_when(or>=or_median_downsampled~1,
#                              TRUE~0),
#          isgreater_uq=case_when(or>=or_uq_downsampled~1,
#                              TRUE~0)
#          )
# pc1pc2_contabs_sims=pc1pc2_contabs_sims%>%
#   mutate(isgreater_raw_median=case_when(or>=or_median_raw~1,
#                              TRUE~0),
#          isgreater_raw_uq=case_when(or>=or_uq_raw~1,
#                              TRUE~0),
#          isgreater_median=case_when(or>=or_median_downsampled~1,
#                              TRUE~0),
#          isgreater_uq=case_when(or>=or_uq_downsampled~1,
#                              TRUE~0)
#          )
# goipc1_contabs_sims=goipc1_contabs_sims%>%
#   mutate(isgreater_raw_median=case_when(or>=or_median_raw~1,
#                              TRUE~0),
#          isgreater_raw_uq=case_when(or>=or_uq_raw~1,
#                              TRUE~0),
#          isgreater_median=case_when(or>=or_median_downsampled~1,
#                              TRUE~0),
#          isgreater_uq=case_when(or>=or_uq_downsampled~1,
#                              TRUE~0)
#          )


pc1rawpc2_contabs_sims=pc1rawpc2_contabs_sims%>%
  mutate(isgreater_raw_median=case_when(or>=or_median_raw~1,
                             TRUE~0),
         isgreater_raw_uq=case_when(or>or_uq_raw~1,
                             TRUE~0),
         isgreater_median=case_when(or>or_median_downsampled~1,
                             TRUE~0),
         isgreater_uq=case_when(or>or_uq_downsampled~1,
                             TRUE~0)
         )
pc1pc2_contabs_sims=pc1pc2_contabs_sims%>%
  mutate(isgreater_raw_median=case_when(or>=or_median_raw~1,
                             TRUE~0),
         isgreater_raw_uq=case_when(or>or_uq_raw~1,
                             TRUE~0),
         isgreater_median=case_when(or>or_median_downsampled~1,
                             TRUE~0),
         isgreater_uq=case_when(or>or_uq_downsampled~1,
                             TRUE~0)
         )
goipc1_contabs_sims=goipc1_contabs_sims%>%
  mutate(isgreater_raw_median=case_when(or>or_median_raw~1,
                             TRUE~0),
         isgreater_raw_uq=case_when(or>or_uq_raw~1,
                             TRUE~0),
         isgreater_median=case_when(or>or_median_downsampled~1,
                             TRUE~0),
         isgreater_uq=case_when(or>or_uq_downsampled~1,
                             TRUE~0)
         )
# pc1pc2_contabs_sims=pc1pc2_contabs_sims%>%
#   mutate(isgreater=case_when(or>=or_pc1pc2~1,
#                              TRUE~0))
# goipc1_contabs_sims=goipc1_contabs_sims%>%
#   mutate(isgreater=case_when(or>=or_pc1pc2~1,
#                              TRUE~0))
# goipc2_contabs_sims=goipc2_contabs_sims%>%
#   mutate(isgreater=case_when(or>=or_pc1pc2~1,
#                              TRUE~0))
pc1rawpc2_isgreater_raw_median=sum(pc1rawpc2_contabs_sims$isgreater_raw_median)
pc1rawpc2_isgreater_raw_uq=sum(pc1rawpc2_contabs_sims$isgreater_raw_uq)
pc1rawpc2_isgreater_median=sum(pc1rawpc2_contabs_sims$isgreater_median)
pc1rawpc2_isgreater_uq=sum(pc1rawpc2_contabs_sims$isgreater_uq)

pc1pc2_isgreater_raw_median=sum(pc1pc2_contabs_sims$isgreater_raw_median)
pc1pc2_isgreater_raw_uq=sum(pc1pc2_contabs_sims$isgreater_raw_uq)
pc1pc2_isgreater_median=sum(pc1pc2_contabs_sims$isgreater_median)
pc1pc2_isgreater_uq=sum(pc1pc2_contabs_sims$isgreater_uq)

goipc1_isgreater_raw_median=sum(goipc1_contabs_sims$isgreater_raw_median)
goipc1_isgreater_raw_uq=sum(goipc1_contabs_sims$isgreater_raw_uq)
goipc1_isgreater_median=sum(goipc1_contabs_sims$isgreater_median)
goipc1_isgreater_uq=sum(goipc1_contabs_sims$isgreater_uq)


# pc1rawpc2_isgreater=sum(pc1rawpc2_contabs_sims$isgreater)
# pc1pc2_isgreater=sum(pc1pc2_contabs_sims$isgreater)
# goipc1_isgreater=sum(goipc1_contabs_sims$isgreater)

simresults=c(cohort_size[k],
             incidence[j],
             or_pair1[i],
             or_pair2[l],
             pc1rawpc2_isgreater_raw_median,
             pc1rawpc2_isgreater_raw_uq,
             pc1rawpc2_isgreater_median,
             pc1rawpc2_isgreater_uq,
             pc1pc2_isgreater_raw_median,
             pc1pc2_isgreater_raw_uq,
             pc1pc2_isgreater_median,
             pc1pc2_isgreater_uq,
             goipc1_isgreater_raw_median,
             goipc1_isgreater_raw_uq,
             goipc1_isgreater_median,
             goipc1_isgreater_uq)

simresults_alldata=c(cohort_size[k],
             incidence[j],
             or_pair1[i],
             or_pair2[l],
    list(goipc1_contabs_sims$or),
    list(pc1pc2_contabs_sims$or))

######Comment out this section to not gather all the 1,000 ORs for each simulation####
simresults_compiled_alldata[[ct]]=simresults_alldata
# a=simresults_compiled_alldata[[1]]

simresults_compiled[ct,]=simresults

ct=ct+1
    }
  }
}
  toc()
}
9.846 sec elapsed
9.035 sec elapsed
9.122 sec elapsed
9.083 sec elapsed
9.072 sec elapsed
9.452 sec elapsed
9.196 sec elapsed
9.349 sec elapsed
9.575 sec elapsed
9.264 sec elapsed
9.143 sec elapsed
9.076 sec elapsed
8.912 sec elapsed
9.049 sec elapsed
9.209 sec elapsed
9.019 sec elapsed
9.234 sec elapsed
simresults_compiled=data.frame(simresults_compiled)
colnames(simresults_compiled)=c("cohort_size",
                                "incidence",
                                "or1",
                                "or2",
                                "pc1rawpc2_isgreater_raw_median",
                                "pc1rawpc2_isgreater_raw_uq",
                                "pc1rawpc2_isgreater_median",
                                "pc1rawpc2_isgreater_uq",
                                "pc1pc2_isgreater_raw_median",
                                "pc1pc2_isgreater_raw_uq",
                                "pc1pc2_isgreater_median",
                                "pc1pc2_isgreater_uq",
                                "goipc1_isgreater_raw_median",
                                "goipc1_isgreater_raw_uq",
                                "goipc1_isgreater_median",
                                "goipc1_isgreater_uq")
simresults_compiled$delta_median=simresults_compiled$goipc1_isgreater_raw_median-simresults_compiled$goipc1_isgreater_median
simresults_compiled$delta_uq=simresults_compiled$goipc1_isgreater_raw_uq-simresults_compiled$goipc1_isgreater_uq

simresults_compiled=simresults_compiled%>%
  mutate(fp_corrected_95=case_when(
    goipc1_isgreater_raw_uq<950~-1,
     (goipc1_isgreater_raw_uq>=950&(goipc1_isgreater_raw_uq-delta_uq)<=950)~1,
    TRUE~0),
    fp_corrected_99=case_when(
      goipc1_isgreater_raw_uq<990~-1,
    (goipc1_isgreater_raw_uq>=990&(goipc1_isgreater_raw_uq-delta_uq)<=990)~1,
    TRUE~0))

# simresults_compiled$goipc1_isgreater_percent=simresults_compiled$goipc1_isgreater*100/1000
simresults_subset=simresults_compiled%>%filter(cohort_size%in%500)

ggplot(simresults_compiled%>%filter(or1<=.4),aes(x=incidence,y=or2))+geom_tile(aes(fill=goipc1_isgreater_median/10))+facet_grid(cohort_size~or1)+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Score")

ggplot(simresults_compiled%>%filter(or1<=.4),aes(x=incidence,y=or2))+geom_tile(aes(fill=goipc1_isgreater_raw_median/10))+facet_grid(cohort_size~or1)+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Percent succesful trials")

ggplot(simresults_compiled%>%filter(or1<=.4),aes(x=incidence,y=or2))+geom_tile(aes(fill=goipc1_isgreater_uq/10))+facet_grid(cohort_size~or1)+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Percent succesful trials")

ggplot(simresults_compiled%>%filter(or1<=.4),aes(x=incidence,y=or2))+geom_tile(aes(fill=goipc1_isgreater_raw_uq/10))+facet_grid(cohort_size~or1)+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Percent succesful trials")

ggplot(simresults_compiled%>%filter(or1<=.7,or1>=.2),aes(x=incidence,y=or2))+geom_tile(aes(fill=delta_median))+facet_grid(cohort_size~or1)+scale_fill_gradient2(low ="red" ,mid ="white",high ="blue",name="Delta bw raw and downsampled")

ggplot(simresults_compiled%>%filter(or1<=.7,or1>=.2),aes(x=incidence,y=or2))+geom_tile(aes(fill=delta_uq))+facet_grid(cohort_size~or1)+scale_fill_gradient2(low ="red" ,mid ="white",high ="blue",name="Delta bw raw and downsampled")

ggplot(simresults_compiled%>%filter(or1<=.7,or1>=.2),aes(x=incidence,y=or2))+geom_tile(color="black",aes(fill=fp_corrected_95))+facet_grid(cohort_size~or1)+scale_fill_gradient2(low ="red" ,mid ="white",high ="blue",name="Regions where downsampling helps")

ggplot(simresults_compiled%>%filter(or1<=.7,or1>=.2),aes(x=incidence,y=or2))+geom_tile(color="black",aes(fill=fp_corrected_99))+facet_grid(cohort_size~or1)+scale_fill_gradient2(low ="red" ,mid ="white",high ="blue",name="Regions where downsampling helps")

ggplot(simresults_compiled%>%filter(or1<=.4),aes(x=incidence,y=or2))+geom_tile(aes(fill=goipc1_isgreater_median/10))+facet_grid(cohort_size~or1)+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Score")

ggplot(simresults_compiled%>%filter(or1<=.4,cohort_size%in%500),aes(x=incidence,y=or2))+geom_tile(aes(fill=goipc1_isgreater_median/10))+facet_wrap(~or1)+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Score")

ggplot(simresults_compiled%>%filter(or1<=.4,cohort_size%in%500),aes(x=incidence,y=or2))+geom_tile(aes(fill=goipc1_isgreater_median/10))+facet_wrap(~or1,ncol=6)+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Score")

ggplot(simresults_compiled%>%filter(or1%in%c(.05,0.1,.5),cohort_size%in%500),aes(x=incidence,y=or2))+geom_tile(color="black",aes(fill=goipc1_isgreater_median/10))+facet_wrap(~or1,ncol=6)+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Score")+scale_x_continuous(expand = c(0,0),name="Incidence")+
  scale_y_continuous(expand = c(0,0),name="Odds ratio of PC1vsPC2")+ggtitle("Odds ratio of GOIvsPC1")+
  theme(plot.title = element_text(hjust = 0.5))

# ggsave("score_heatmap.pdf",width=8,heigh=3,units="in",useDingbats=F)

ggplot(simresults_compiled%>%filter(or1%in%c(.05,0.1,.5),cohort_size%in%c(100,500,900)),aes(x=incidence,y=or2))+geom_tile(color="black",aes(fill=goipc1_isgreater_median/10))+facet_grid(cohort_size~or1)+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Score")+scale_x_continuous(expand = c(0,0),name="Incidence")+
  scale_y_continuous(expand = c(0,0),name="Odds ratio of PC1vsPC2")+ggtitle("Odds ratio of GOIvsPC1")+
  theme(plot.title = element_text(hjust = 0.5))

# ggsave("score_heatmap_supplement.pdf",width=8,heigh=7,units="in",useDingbats=F)


ggplot(simresults_compiled%>%filter(or2%in%.05,cohort_size%in%c(500)),aes(x=factor(or1),y=goipc1_isgreater_median/10))+geom_boxplot()+scale_x_discrete(name="Odds ratio of PC1vsPC2")+
  scale_y_continuous(name="Score")+
  theme(plot.title = element_text(hjust = 0.5))+cleanup

# ggsave("score_plot.pdf",width=8,heigh=3,units="in",useDingbats=F)



ggplot(simresults_compiled%>%filter(incidence%in%c(6,16,26,36),cohort_size%in%c(100,300,500,900)),aes(x=factor(or1),y=or2))+geom_tile(color="black",aes(fill=goipc1_isgreater_uq/10))+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Score")+scale_x_discrete(expand = c(0,0),name="Odds ratio of GOI and PC1")+facet_grid(cohort_size~incidence)+
  scale_y_continuous(expand = c(0,0),name="Odds ratio of PC1 and PC2")+theme(plot.title = element_text(hjust = 0.5))

ggplot(simresults_compiled%>%filter(incidence%in%c(6,16,26,36),cohort_size%in%500),aes(x=factor(or1),y=or2))+geom_tile(color="black",aes(fill=goipc1_isgreater_uq/10))+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Score")+scale_x_discrete(expand = c(0,0),name="Odds ratio of GOI and PC1")+facet_wrap(~incidence,ncol=4)+
  scale_y_continuous(expand = c(0,0),name="Odds ratio of PC1 and PC2")+theme(plot.title = element_text(hjust = 0.5),axis.text.x=element_text(angle=90,hjust=.5,vjust=.5),axis.text=element_text(face="bold",size="9",color="black"))

# ggsave("score_heatmap_bestoption1.pdf",width=8,heigh=2.5,units="in",useDingbats=F)
# sort(unique(simresults_compiled$or2))

ggplot(simresults_compiled%>%filter(cohort_size%in%500),aes(x=factor(or1),y=or2))+geom_tile(color="black",aes(fill=goipc1_isgreater_median/10))+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Score")+scale_x_discrete(expand = c(0,0),name="Odds ratio of GOI and PC1")+
  scale_y_continuous(expand = c(0,0),name="Odds ratio of PC1 and PC2")+theme(plot.title = element_text(hjust = 0.5))

# ggsave("score_heatmap_bestoption.pdf",width=4,heigh=3,units="in",useDingbats=F)

ggplot(simresults_compiled%>%filter(or1%in%.3,or2%in%.09),aes(x=incidence/cohort_size,y=goipc1_isgreater_median/10))+geom_point()

ggplot(simresults_compiled%>%filter(or1%in%.1,or2%in%.05,cohort_size%in%500),aes(x=incidence,y=goipc1_isgreater_median/10))+geom_point()

  # scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Score")+scale_x_discrete(expand = c(0,0),name="Odds ratio of GOI and PC1")+
  # scale_y_continuous(expand = c(0,0),name="Odds ratio of PC1 and PC2")+theme(plot.title = element_text(hjust = 0.5))
# or1%in%c(.05,0.1,.5),cohort_size%in%c(100,500,900)
ggplot(simresults_compiled%>%filter(cohort_size%in%c(100,300,500)),aes(x=incidence,y=or2))+geom_tile(color="black",aes(fill=goipc1_isgreater_median/10))+facet_grid(cohort_size~or1)+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Score")+scale_x_continuous(expand = c(0,0),name="Incidence")+
  scale_y_continuous(expand = c(0,0),name="Odds ratio of PC1vsPC2")+ggtitle("Odds ratio of GOIvsPC1")+
  theme(plot.title = element_text(hjust = 0.5))

Choosing plots for putting in paper

ggplot(simresults_compiled%>%filter(cohort_size%in%500),aes(x=factor(or1),y=or2))+geom_tile(color="black",aes(fill=goipc1_isgreater_uq/10))+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Score")+scale_x_discrete(expand = c(0,0),name="Odds ratio of GOI and PC1")+
  scale_y_continuous(expand = c(0,0),name="Odds ratio of PC1 and PC2")+theme(plot.title = element_text(hjust = 0.5))+theme(plot.title = element_text(hjust = 0.5),axis.text.x=element_text(angle=90,hjust=.5,vjust=.5),axis.text=element_text(face="bold",size="9",color="black"))

# ggsave("score_heatmap_bestoption.pdf",width=4,heigh=3,units="in",useDingbats=F)

ggplot(simresults_compiled%>%filter(or1%in%c(.05,.3,.6),incidence%in%c(4,8,12,16,20,24,28,32,36),cohort_size%in%500),aes(x=factor(incidence),y=or2))+geom_tile(color="black",aes(fill=goipc1_isgreater_median/10))+facet_wrap(~or1,ncol=6)+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Score")+scale_x_discrete(expand = c(0,0),name="Incidence")+
  scale_y_continuous(expand = c(0,0),name="Odds ratio of PC1 and PC2")+theme(strip.text=element_blank(),plot.title = element_text(hjust = 0.5),axis.text.x=element_text(angle=90,hjust=.5,vjust=.5),axis.text=element_text(face="bold",size="9",color="black"),legend.position = "none")

# ggsave("score_heatmap_bestoption1.pdf",width=6,heigh=2.5,units="in",useDingbats=F)

ggplot(simresults_compiled%>%filter(or1%in%c(.1,.2,.5),incidence%in%c(4,8,12,16,20,24,28,32,36),cohort_size%in%500),aes(x=factor(incidence),y=or2))+geom_tile(color="black",aes(fill=goipc1_isgreater_raw_median/10))+facet_wrap(~or1,ncol=6)+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Score")+scale_x_discrete(expand = c(0,0),name="Incidence")+
  scale_y_continuous(expand = c(0,0),name="Odds ratio of PC1 and PC2")+theme(strip.text=element_blank(),plot.title = element_text(hjust = 0.5),axis.text.x=element_text(angle=90,hjust=.5,vjust=.5),axis.text=element_text(face="bold",size="9",color="black"),legend.position = "none")

# ggsave("score_heatmap_bestoption1.pdf",width=6,heigh=2.5,units="in",useDingbats=F)

ggplot(simresults_compiled%>%filter(cohort_size%in%500,incidence%in%c(4,8,12,16,20,24,28,32,36)),aes(x=factor(incidence),y=or2))+geom_tile(color="black",aes(fill=goipc1_isgreater_raw_median/10))+facet_wrap(~or1,ncol=4)+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Score")+scale_x_discrete(expand = c(0,0),name="Incidence")+
  scale_y_continuous(expand = c(0,0),name="Odds ratio of PC1 and PC2")+theme(plot.title = element_text(hjust = 0.5),axis.text.x=element_text(angle=90,hjust=.5,vjust=.5),axis.text=element_text(face="bold",size="9",color="black"),legend.position = "none")

# ggsave("score_heatmap_bestoption1_supplement.pdf",width=6,heigh=6,units="in",useDingbats=F)

ggplot(simresults_compiled%>%filter(or1%in%c(.05,.1,.5),cohort_size%in%c(100,300,500)),aes(x=incidence,y=or2))+geom_tile(color="black",aes(fill=goipc1_isgreater_median/10))+facet_grid(cohort_size~or1)+scale_fill_gradient2(low ="red" ,mid ="white",midpoint=50,high ="blue",name="Score")+scale_x_discrete(expand = c(0,0),name="Incidence")+
  scale_y_continuous(expand = c(0,0),name="Odds ratio of PC1 and PC2")

# a=simresults_compiled%>%filter(cohort_size%in%500,or1%in%.7,or2%in%.05,incidence%in%)
# simresults_subset=simresults_compiled%>%filter(or2%in%0.01,or1%in%c(0.01,.1,1))
simresults_subset=unlist(simresults_compiled_alldata)

simresults_unlisted=data.frame(unlist(lapply(simresults_compiled_alldata,'[[',1)),
             unlist(lapply(simresults_compiled_alldata,'[[',2)),
             unlist(lapply(simresults_compiled_alldata,'[[',3)),
             unlist(lapply(simresults_compiled_alldata,'[[',4)))
simresults_unlisted$list1=(lapply(simresults_compiled_alldata,'[[',5))
simresults_unlisted$list2=(lapply(simresults_compiled_alldata,'[[',6))

colnames(simresults_unlisted)=c("cohort_size","incidence","or1","or2","or1_list","or2_list")

simresults_unlisted=simresults_unlisted%>%filter(or2%in%0.01,or1%in%c(0.05,.1,1),cohort_size%in%500,incidence%in%c(4,8,12,16,20))
# library(reshape2)
a=simresults_unlisted%>%filter(incidence%in%8,or2%in%0.01,or1%in%.05)
# b=unnest(a)
median(b$or2_list)>median(b$or1_list)

library(tidyr)
simresults_melted=unnest(simresults_unlisted)
simresults_melted2=melt(simresults_melted,
                        id.vars = c("cohort_size","or1","or2","incidence"),
                        measure.vars =c("or1_list","or2_list"),
                        variable.name = "Comparison",
                        value.name = "OR"
                        )
ggplot(simresults_melted2,aes(x=factor(incidence),y=OR,fill=Comparison))+facet_wrap(~or1)+geom_boxplot()+scale_y_continuous(trans="log2")+cleanup

plotly=ggplot(simresults_melted2,aes(x=factor(incidence),y=OR,fill=Comparison))+facet_wrap(~or1)+geom_boxplot()+cleanup
ggplotly(plotly)










simresults_unlisted=data.frame(unlist(lapply(simresults_compiled_alldata,'[[',1)),
             unlist(lapply(simresults_compiled_alldata,'[[',2)),
             unlist(lapply(simresults_compiled_alldata,'[[',3)),
             unlist(lapply(simresults_compiled_alldata,'[[',4)))
simresults_unlisted$list1=(lapply(simresults_compiled_alldata,'[[',5))
simresults_unlisted$list2=(lapply(simresults_compiled_alldata,'[[',6))

colnames(simresults_unlisted)=c("cohort_size","incidence","or1","or2","or1_list","or2_list")
simresults_unlisted=simresults_unlisted%>%filter(or1%in%.5,or2%in%c(0.05,.11,.19),cohort_size%in%500,incidence%in%c(4,8,12,16,20))

simresults_melted=unnest(simresults_unlisted)
simresults_melted2=melt(simresults_melted,
                        id.vars = c("cohort_size","or1","or2","incidence"),
                        measure.vars =c("or1_list","or2_list"),
                        variable.name = "Comparison",
                        value.name = "OR"
                        )

ggplot(simresults_melted2,aes(x=factor(incidence),y=OR,fill=Comparison))+facet_wrap(~or2)+geom_boxplot()+scale_y_continuous(trans="log2")+cleanup

plotly=ggplot(simresults_melted2,aes(x=factor(incidence),y=OR,fill=Comparison))+facet_wrap(~or2)+geom_boxplot()+cleanup
ggplotly(plotly)
# sort(unique(simresults_unlisted$or2))

#I used the following website to make sure that my Gaussian elimination was working properly
# https://www.emathhelp.net/calculators/linear-algebra/gauss-jordan-elimination-calculator/?i=%5B%5B1%2C1%2C0%2C0%2C.4836%5D%2C%5B1%2C1%2C1%2C1%2C1%5D%2C%5B0%2C1%2C0%2C-1%2C0%5D%2C%5B1%2C0%2C-0.01%2C0%2C0%5D%5D&reduced=on

sessionInfo()
R version 4.0.0 (2020-04-24)
Platform: x86_64-apple-darwin17.0 (64-bit)
Running under: macOS  10.16

Matrix products: default
BLAS:   /Library/Frameworks/R.framework/Versions/4.0/Resources/lib/libRblas.dylib
LAPACK: /Library/Frameworks/R.framework/Versions/4.0/Resources/lib/libRlapack.dylib

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] parallel  grid      stats     graphics  grDevices utils     datasets 
[8] methods   base     

other attached packages:
 [1] BiocManager_1.30.10 plotly_4.9.2.1      ggsignif_0.6.0     
 [4] devtools_2.3.0      usethis_1.6.1       RColorBrewer_1.1-2 
 [7] reshape2_1.4.4      ggplot2_3.3.3       doParallel_1.0.15  
[10] iterators_1.0.12    foreach_1.5.0       dplyr_1.0.6        
[13] VennDiagram_1.6.20  futile.logger_1.4.3 tictoc_1.0         
[16] knitr_1.28          workflowr_1.6.2    

loaded via a namespace (and not attached):
 [1] Rcpp_1.0.4.6         tidyr_1.1.3          prettyunits_1.1.1   
 [4] ps_1.3.3             assertthat_0.2.1     rprojroot_1.3-2     
 [7] digest_0.6.25        utf8_1.1.4           R6_2.4.1            
[10] plyr_1.8.6           futile.options_1.0.1 backports_1.1.7     
[13] evaluate_0.14        httr_1.4.2           pillar_1.6.1        
[16] rlang_0.4.11         lazyeval_0.2.2       data.table_1.12.8   
[19] whisker_0.4          callr_3.7.0          rmarkdown_2.8       
[22] labeling_0.3         desc_1.2.0           stringr_1.4.0       
[25] htmlwidgets_1.5.1    munsell_0.5.0        compiler_4.0.0      
[28] httpuv_1.5.2         xfun_0.22            pkgconfig_2.0.3     
[31] pkgbuild_1.0.8       htmltools_0.4.0      tidyselect_1.1.0    
[34] tibble_3.1.2         codetools_0.2-16     viridisLite_0.3.0   
[37] fansi_0.4.1          crayon_1.4.1         withr_2.4.2         
[40] later_1.0.0          jsonlite_1.7.2       gtable_0.3.0        
[43] lifecycle_1.0.0      DBI_1.1.0            git2r_0.27.1        
[46] magrittr_2.0.1       formatR_1.7          scales_1.1.1        
[49] cli_2.5.0            stringi_1.4.6        farver_2.0.3        
[52] fs_1.4.1             promises_1.1.0       remotes_2.1.1       
[55] testthat_2.3.2       ellipsis_0.3.2       generics_0.0.2      
[58] vctrs_0.3.8          lambda.r_1.2.4       tools_4.0.0         
[61] glue_1.4.1           purrr_0.3.4          processx_3.5.2      
[64] pkgload_1.0.2        yaml_2.2.1           colorspace_1.4-1    
[67] sessioninfo_1.1.1    memoise_1.1.0