1-Preprocessing.Rmd

---
title: "Data Collection and Pre-processing"
author: Chuleeporn Phanus-umporn, Saw Simeon and Chanin Nantasenamat
date: "August 22, 2016"
output: word_document
---

## Import data

Import the bioactivity data obtained from the ChEMBL database as a dataframe. Select the IC50 data subset.

```{r, warning = FALSE, message = FALSE, error = FALSE, tidy = TRUE}
library(readxl)
df_raw <- suppressWarnings(read_excel("Human-HDAC1-ChEMBL21.xlsx"))
names <- c("CMPD_CHEMBLID", "CANONICAL_SMILES", "STANDARD_TYPE",
           "RELATION", "STANDARD_VALUE", "STANDARD_UNITS",
           "PROTEIN_ACCESSION", "PREF_NAME", 
      "PUBMED_ID",
      "JOURNAL", "YEAR", "VOLUME", 
      "ISSUE", "FIRST_PAGE", "MOLWEIGHT", 
      "ALOGP", 
           "PSA", "NUM_RO5_VIOLATIONS")
df <- df_raw[, names]
df_1 <- subset(df, STANDARD_TYPE == "IC50")
df_2 <- subset(df_1, RELATION == "=")
df_4 <- subset(df_2, STANDARD_UNITS == "nM")

bioactivity <- paste0("Bioactivity data points: ", nrow(df))
IC50 <- paste0("IC50 Bioactivity data points: ", nrow(df_1))
IC50_omit <- paste0("Omit IC50 with lesser/greater than symbol: ", nrow(df_2))
final <- paste0("Final number of data points: ", nrow(df_4))

message <- c(bioactivity, IC50, IC50_omit, final)
print(message)
```

## Remove redundant rows

```{r, tidy= TRUE, warning = FALSE, message = FALSE, error = FALSE}

ID <- df_4$CMPD_CHEMBLID
IC50_nm <- df_4$STANDARD_VALUE
smiles <- df_4$CANONICAL_SMILES
data_1 <- data.frame(ID, smiles,  IC50_nm)

duplicate <- smiles[duplicated(smiles)]
unique <- unique(duplicate)
unique <- as.character(unique)
results <- data.frame()
for (i in unique) {
  pre_data <- subset(data_1, smiles == i)
  results <- rbind(results, pre_data)
}


results_2 <- data.frame()
for (i in unique) {
  class <- subset(results, smiles == i)
  IC50 <- class$IC50_nm
  mean_IC50_nm <- mean(IC50)
  sd <- sd(IC50)
  data_frame <- data.frame(i, mean_IC50_nm, sd)
  results_2 <- rbind(data_frame, results_2)
}


keep <- subset(results_2, sd < 2)
names(keep) <- c("CANONICAL_SMILES", "STANDARD_VALUE", "SD")


non_redundant <- df_4[!duplicated(df_4$CANONICAL_SMILES), ]
non_redundant <- non_redundant[!non_redundant$CANONICAL_SMILES %in% unique, ] 

post_df_1 <- subset(non_redundant, RELATION == "=")
post_df_2 <- subset(post_df_1, STANDARD_UNITS == "nM")


non_redundant <- post_df_2[, c("CANONICAL_SMILES", "STANDARD_VALUE")]
redundant <- df_4[df_4$CANONICAL_SMILES %in% unique, ]  #########################
redundant <- subset(redundant, !duplicated(redundant[, c("CANONICAL_SMILES")]))
redundant <- subset(redundant, select = -STANDARD_VALUE)
STANDARD_VALUE <- keep$STANDARD_VALUE
CANONICAL_SMILES <- keep$CANONICAL_SMILES
cleaned_redundant <- data.frame(CANONICAL_SMILES, STANDARD_VALUE)
curated_data <- rbind(non_redundant, cleaned_redundant)
curated_data_na_removed <- na.omit(curated_data)

non_redundant_compound <- paste0("number of non-redundant compounds: ", 
                                 nrow(non_redundant))
unique <- paste0("number of unique redundant compounds: ", length(unique))
sd_2 <- paste0("Compounds left that are kept", " which has a sd of lower than two: ",
               nrow(keep))
combined <- paste0("Combined data points: ", nrow(curated_data))
curated <- paste0("number of curated data points where missing smiles are removed: ", nrow(curated_data_na_removed))

message <- c(non_redundant_compound, unique, sd_2, combined, curated)
print(message)

```


## Save the resulting data frame as an Excel file

```{r, tidy = TRUE, warning = FALSE, message = FALSE, error = FALSE}

library(xlsx)

curated_smiles <- curated_data_na_removed$CANONICAL_SMILES
STANDARD_VALUE <- curated_data_na_removed$STANDARD_VALUE
new_df <- df_4[df_4$CANONICAL_SMILES %in% curated_smiles, ]
get_unique <- new_df[!duplicated(new_df$CANONICAL_SMILES), ]
#new_df_2 <- get_unique[, !(names(get_unique) %in% "STANDARD_TYPE")]

write.xlsx(get_unique, file = "pre-processed.xlsx", sheetName = "Sheet1", row.names = FALSE)
```

## Create a SMILES file

```{r, tidy = TRUE, warning = FALSE, message = FALSE, error = FALSE}

library(xlsx)
library(readxl)

# df_preprocessed <- suppressWarnings(read_excel("pre-processed.xlsx"))
cpd_smiles <- get_unique[, c("CANONICAL_SMILES", "CMPD_CHEMBLID")]

# fileConn<-file("compounds.smi")
# writeLines(cpd_smiles, fileConn)
# close(fileConn)

write.csv(cpd_smiles, file = "compounds.smi", quote = FALSE, row.names = FALSE, col.names = FALSE)
```