dbGaPdb_app.Rmd

---
title: "Selecting dbGaP studies using metadata queries"
resource_files:
- demo_dbgap_metadata.Rdata
runtime: shiny
output:
  flexdashboard::flex_dashboard:
    social: menu
---

Results
============================================


Column {.sidebar}
-----------------------------------------------------------------------

  
```{r}


#Load the proper libraries
Sys.setenv(RSTUDIO_PANDOC="/usr/lib/rstudio/bin/pandoc")
enableBookmarking()
library("tm")
library("tidyverse")
library("shiny")
library("DT")
library("RSQLite")
library("dbplyr")
library(wordcloud)
#Load local data
#load("demo_dbgap_metadata.Rdata")

#Collect user input to later subset the data
inputPanel(

  selectizeInput("dat", label = h4("Data Type"), c("", "RNA-seq", "Whole Genome Sequencing", "Whole Exome Sequencing")),

  selectizeInput("pheno", label = h4("Phenotype"), c("", "asthma", "insomnia", "hypertension")),
  
  selectizeInput("n", label = h4("Minimum sample size"), c(0,10,50,100,500,1000,5000,10000), selected = 10000)
  
)
dbGaPdb <- src_sqlite('dbGaPdb.sqlite')
study_variable_info_SQL <<- tbl(dbGaPdb, "study_variable_info")

```


Column
-----------------------------------------------------------------------

Studies containing your queries:

```{r}


#Make a subset of the data based on the search terms
subset_dat <- reactive({

inph <- paste0("%",input$pheno,"%")
indt <- paste0("%",input$dat,"%")
inn <- input$n

#  tables$study_variable_info[(as.numeric(tables$study_variable_info$male_count) + as.numeric(tables$study_variable_info$female_count) > as.numeric(input$n)),] %>% filter(grepl(input$pheno, description, ignore.case=T)&grepl(input$dat ,description, ignore.case=T))
  
study_variable_info_SQL %>% mutate(mf_count = as.numeric(male_count) + as.numeric(female_count)) %>% filter(mf_count > as.numeric(inn)) %>% filter(description %LIKE% inph) %>% filter(description %LIKE% indt) %>% data.frame()
  
  #study_variable_info_SQL %>% mutate(mf_count = as.numeric(male_count) + as.numeric(female_count)) %>% filter(mf_count > 10000) %>% filter(grepl("asthma", description, ignore.case=T)&grepl("" ,description, ignore.case=T)) %>% head() %>% data.frame()
   
})


#Render a data table of the subsetted data
renderDataTable({

  subset_dat()
  
})


```


Word Cloud
========================================


```{r}

#study_dataset_info_SQL <- tbl(dbGaPdb, "study_dataset_info")

df_sdis <- tbl(dbGaPdb, 'study_dataset_info') %>% select(study_accession, description) %>% data.frame()


out_ccat <- reactive({

inph <- paste0("%",input$pheno,"%")
indt <- paste0("%",input$dat,"%")
inn <- input$n
  
to_filt <- study_variable_info_SQL %>% mutate(mf_count = as.numeric(male_count) + as.numeric(female_count)) %>% filter(mf_count > as.numeric(inn)) %>% filter(description %LIKE% inph) %>% filter(description %LIKE% indt) %>% data.frame()

vec_to_filt <- to_filt$study_accession

filt <- df_sdis[df_sdis$study_accession %in% vec_to_filt,]


ccat <- ""

for (each in filt$description){
  ccat <- paste0(ccat, each)
}

new_ccat <- gsub('quot','',ccat)
jeopCorpus <- Corpus(VectorSource(new_ccat))
jeopCorpus <- tm_map(jeopCorpus, PlainTextDocument)
jeopCorpus <- tm_map(jeopCorpus, removeWords, stopwords('english'))

})


renderPlot({
  
  wordcloud(out_ccat(), min.freq = 5, max.words = 100, colors=brewer.pal(6,"Dark2"), scale=c(6,2), random.order = FALSE)
  
})

```



Download
========================================


Column
-----------------------------------------------------------------------




```{r}


#Download page
downloadHandler(filename = function(){
  ("dbGaPdb_query.csv")},
  content = function(file) {
    write.csv(subset_dat(), file)
  }

)


```