Homework2

knitr::opts_chunk$set(echo = TRUE)

In this homework we will write a function to read in and summarize data files in the format of the example file hmp.txt, which you can obtain from canvas or GitHub. When string manipulation is necessary, please use functions from stringr.

library(stringr)

1. Determine the delimiter of the file hmp.txt and use the appropriate read. function to read it into an R data frame named hmp. Do not read the character strings in as factors. Save the dimensions of hmp in a variable hmp.dim.

setwd("~/Desktop/课业文档/Statistics/STAT 341/SFUStat341/Homeworks")
hmp<-read.delim("hmp.txt",header=TRUE,sep="\t",stringsAsFactors = FALSE)
str(hmp)
hmp.dim<-dim(hmp)

2. Extract the first 11 columns of hmp, and save them as a data frame called marker.info. In the ninth column of marker.info you will see long character strings. Some of these contain the substring MITOCHONDRIA. Create a logical vector mit that is TRUE for strings that contain MITOCHONDRIA and FALSE otherwise. Add mit to the marker.info data frame.

marker.info<-data.frame(hmp[,1:11])
str(marker.info)
marker.info$mit<-regexpr("MITOCHONDRIA",marker.info[,9])>0
marker.info

3. Progamatically extract the remaining columns (that is, write code that uses hmp.dim to calculate the indices of the columns to extract), transpose this subset and save as a matrix marker.matrix. (Notice that transposing has coerced marker.matrix to a matrix of characters and that the column names from hmp have become the row names of marker.matrix.) Set the column names of marker.matrix to be the character strings in the first column of marker.info (strings that start with rs). Print the first six rows of marker.matrix.

4. Background: The NN entries in marker.matrix are missing data. Other than missing data, the possible letters in a column of marker.matrix are given in the alleles column of the corresponding row of marker.info. For example, the possible letters in the first column of marker.matrix are A and T. The "minor" allele is the letter that is least common in the data, once missing data are excluded. Hint: If vv is a column of marker.matrix, then str_split_fixed(vv,pattern="",n=2) would be useful (str_split_fixed is from stringr). Your task: Write a function called MAsummary() that takes a column of marker.matrix and the corresponding alleles entry from model.info as input and returns a list with elements (i) MA, the minor allele (the letter), (ii) MAF, the frequency of the minor allele (proportion of letters other than N that are the MA), and (iii) MAcount, a vector of minor allele counts for each entry of the input vector. Test your function on the first column of marker.matrix with MAsummary(marker.matrix[,1],marker.info[1,"alleles"]) Your function should find that the MA is A, the MAF is 0.0227, and the first six elements of MAcount are c(NA,0,NA,NA,0,0).

5. Create vectors MA and MAF of length nrow(marker.info) and a matrix MAcount with dimension dim(marker.matrix). Use a for loop with index i to loop through the columns of marker.matrix and use the output of MAsummary(marker.matrix[,i],marker.info[i,"alleles"]) to fill in the ith elements of MA and MAF, and the ith column of MAcount. Add MA and MAF to the marker.info data frame. Copy MAcount to marker.matrix; that is, overwrite marker.matrix with MAcount.

6. Verify your code by (i) bundling marker.info and MAcount into a list with mylist <- list(marker.info=marker.info,MAcount=MAcount), and (ii) using identical() to compare mylist to the output in the object testlist that you can load from [http://people.stat.sfu.ca/~mcneney/Teaching/Stat341/Test/testhw2.rda]

setwd("~/Desktop/课业文档/Statistics/STAT 341/SFUStat341/Homeworks") hmp<-read.delim("hmp.txt",header=TRUE,sep="\t",stringsAsFactors = FALSE) str(hmp) hmp.dim<-dim(hmp)