Department of Genetics

Luiz de Queiroz College of Agriculture

University of São Paulo

R is a language and environment for statistical computing and graphics. To download R, please visit the Comprehensive R Archive Network. You do not need to be an expert on it to be able to build your linkage map using *OneMap*.

Although we prefer and recommend the Linux version, in this tutorial it is assumed that the user is running Windows. Users of R under Linux or Mac OS should have no difficulty following this tutorial.

We would like to recommend that new users, instead of using plain R, use it through the fantastic software RStudio. With this package, there is no noticeable difference between operating systems.

As advertised on the website, *RStudio is an integrated development environment (IDE) for R. It includes a console, syntax-highlighting editor that supports direct code execution, as well as tools for plotting, history, debugging and workspace management*. In other words, it offers a number of facilities for your convenience that will make your life easier, specially if you have never used R before.

So, go ahead and download and install R and RStudio. The window on the left is where you type the R commands you want.

In the left window, you can see a *greater than* sign (``>’’), which means that R is waiting for a command. We call this a *prompt*.

Let us start with a simple example adding two numbers. Type `2 + 3`

at the prompt then type the *Enter* key. You will see the result directly on the screen.

You can store this result into a variable for future use, applying the assignment operator _ <- _ (*less than* sign and _ minus_ altogether):

The result of the calculation was stored into the variable *x*. You can access this result typing *x* at the prompt:

You can also use the variable *x* into other calculations. For example:

So, play a little just to start understanding what is going on.

Another fundamental aspect in R is the usage of *functions*. A function is a predefined routine used to do specific calculations. For example, to calculate the natural logarithm of \(6.7\), we can use the function *log*:

The function *log* contains a group of internal procedures to calculate the natural logarithm of a positive real number. The input values of a function are called *arguments*.

In the previous example, we provided only one argument (\(6.7\)) to the function. Sometimes a function has more than one argument. For example, to obtain the logarithm of \(6.7\) to base \(4\), you can use:

It is possible to calculate the natural logarithm of a set of numbers by defining a vector and using it as the first argument of the function *log*. To do so we use the function *c()*, that *combines* a set of values into a vector. Thus, to calculate the logarithm of the numbers 6.7, 3.2, 5.4, 8.1, 4.9, 9.7 and 2.5, one can use:

```
y <- c(6.7, 3.2, 5.4, 8.1, 4.9, 9.7, 2.5)
log(y)
#> [1] 1.9021075 1.1631508 1.6863990 2.0918641 1.5892352 2.2721259 0.9162907
```

Notice that *y* is a vector, that is the argument to the function *log()*.

Every R function has a help page which can be accessed using a question mark before the name of the function. For example, to get help on function *log*, you would type:

This command will open a help page in the default web browser of your system. The help page contains some important information about the function such its syntax, its arguments and some usage examples.

There are many other ways of getting help, of course. For example, from RStudio, click *Help* on the menu. For doing searches on the internet, it is better to first go to http://rseek.org/, since R is a very common letter to include in searches.

Although R has a huge amount of internal functions, for doing very specific computations (like constructing genetic linkage maps), it is necessary to add extra functionalities. These can be done by installing a *package* (that, loosely speaking, will include a number of new functions for helping you to achieve what you are trying to do). A package is a collection of related functions, help files and example data files that have been bundled together (Adler, 2010).

For example, let us assume that you need to convert a set of recombination fractions into centimorgan distance using the Kosambi mapping function. One possible way to do this is by using basic R to write a function to calculate the distances. Another way is use the *OneMap* package. To install it you can type:

You also can use the console menus on RStudio. On the bottom window to the right, select **Packages**, then **Install** and finally select *OneMap* (select CRAN as your repository). Yes, it is that easy!

Returning to the console, you need to load *OneMap* by typing:

Some Linux users reported the error message below:

To fix it, in a terminal (outside R), install `r-cran-tkrplot`

:

To finish our example, let us enter some recombination fractions, for example, 0.01, 0.12, 0.05, 0.11, 0.21, 0.07, and save it into a variable named *rf*:

Now, let us use *OneMap*’s function *kosambi* to do the calculation:

You can also obtain help on the function *kosambi* using the question mark in the same way as done before:

So far, we entered the variables in R by typing them directly into the console. However, in real situations we usually **read these values from a file** or a data bank (including files on the internet).

To learn this procedure, copy and paste the following table into a text editor (for example, *notepad*) and save it to a file called *test.txt* into any directory in your computer (such as *My Documents*).

```
x y
2.13 4.50
4.48 1.98
10.95 9.29
10.03 16.25
12.72 27.38
24.63 22.60
22.57 36.87
29.78 31.73
19.54 10.42
7.86 14.68
11.75 8.68
23.71 37.39
```

To read these data set into R, first, you have to set the working directory. Go to *Session*, then *Set Working Directory*, and *Choose Directory*, pointing to where you saved the file *test.txt*.

Now let us read the file *test.txt* into R and store it in a variable named *dat*. To do this, we can use using the R function *read.table*. The first argument is the name of the file; the second one indicates if the file contains a header, that is, if the first line of the file contains the names of the variables (which is true for our example):

The second line, with *dat*, is necessary to ask R to print the contents of the object *dat* (i. e., the data itself). Inspecting the object *dat* you can see a table with 12 rows and two columns. The names of the columns are *x* and *y*. We can access the variables in columns using the dollar sign followed by the column name:

```
dat$x
#> [1] 2.13 4.48 10.95 10.03 12.72 24.63 22.57 29.78 19.54 7.86 11.75
#> [12] 23.71
dat$y
#> [1] 4.50 1.98 9.29 16.25 27.38 22.60 36.87 31.73 10.42 14.68 8.68
#> [12] 37.39
```

It is also possible to use the function *summary* to extract some information about the object *dat*, or about each one of the columns separately:

```
summary(dat)
#> x y
#> Min. : 2.130 Min. : 1.980
#> 1st Qu.: 9.488 1st Qu.: 9.137
#> Median :12.235 Median :15.465
#> Mean :15.012 Mean :18.481
#> 3rd Qu.:22.855 3rd Qu.:28.468
#> Max. :29.780 Max. :37.390
summary(dat$x)
#> Min. 1st Qu. Median Mean 3rd Qu. Max.
#> 2.130 9.488 12.235 15.012 22.855 29.780
summary(dat$y)
#> Min. 1st Qu. Median Mean 3rd Qu. Max.
#> 1.980 9.137 15.465 18.481 28.468 37.390
```

The function *summary* provides some descriptive statistics about the variables in the dataset. If you want to export these information to a file you can use the function *write.table*:

The first argument is the output of the *summary* function. Note that is possible to use a function as an argument of another one. The second argument is the name of the file in which the summary will be written. Notice that this will happen in the *working directory*, previously set through RStudio menus. The third argument eliminates double quotes from the output file. After running the command, you can look for the file *test_sum.txt* in the working directory you defined before.

In R, every object belongs to a ** class**. This is a simple concept that you must remember. For example, the

When we used the function *summary*, it automatically recognized the class of the object *dat* and applied a specific procedure developed for class *data.frame*, which in this case involves the computation of some descriptive statistics.

This procedure is named *method*. However, other classes of objects can be used as arguments to function *summary* and the result will be different!

For example, let us adjust a linear (regression) model using column *y* as the response variable, and column *x* as the independent one. This can be done with the function *lm()*:

```
ft_mod <- lm(dat$y ~ dat$x)
ft_mod
#>
#> Call:
#> lm(formula = dat$y ~ dat$x)
#>
#> Coefficients:
#> (Intercept) dat$x
#> 1.803 1.111
```

This function is used to fit linear models and, by default, prints just a formula and the coefficients of the linear regression. Object *ft_mod* is of class *lm*:

So, if we use function *summary* to obtain more information about the fitted model, the result will be:

```
summary(ft_mod)
#>
#> Call:
#> lm(formula = dat$y ~ dat$x)
#>
#> Residuals:
#> Min 1Q Median 3Q Max
#> -13.091 -5.144 -1.413 5.421 11.446
#>
#> Coefficients:
#> Estimate Std. Error t value Pr(>|t|)
#> (Intercept) 1.8026 4.7689 0.378 0.71334
#> dat$x 1.1110 0.2771 4.009 0.00248 **
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> Residual standard error: 8.075 on 10 degrees of freedom
#> Multiple R-squared: 0.6164, Adjusted R-squared: 0.5781
#> F-statistic: 16.07 on 1 and 10 DF, p-value: 0.002482
```

In this case, function *summary* recognizes *ft_mod* as an object of class *lm* and applies a method which shows information about the fitted model, such as the distribution of the residuals, regression coefficients, t-tests, and the coefficient of determination (\(R^2\)), etc.

Thus, it is possible to use the same function on different classes of object to obtain different results. This concept is very important in *OneMap* and you must remember it to use the package. For example, in other vignettes, we will show that depending on the class of the dataset, which can be *outcross*, *f2*, *backcross*, *riself* and *risib*, a certain set of procedures will be applied. Not by coincidence, these classes correspond to all types of populations that can be analyzed. The advantage of this approach is that you do not need to change the function to do a specific analysis; it will recognize the object type and will adapt accordingly.

Finally, you may need to save your work to come back to it in another working session. But before we explain how to do that, let us explain a few other concepts.

You can save your ** R Script**, which is the file that has all R instructions you typed so far. You can later load them and run all instructions again to get the same results. This is easy: just click

A different thing is to save your **R Session**, with all objects you created so far (called *R Workspace*). This is not the same, because once you load the workspace, you will have all the objects already loaded, not requiring you to do everything again, i. e, running your script. This will help you to save a lot of time, since some of the analysis required to build linkage maps are time demanding.

To do so, click *Session*, then *Save Workspace As* and choose a directory and name. In your next session, open RStudio and then go to *Session*, *Load Workspace*.

Alternatively, you can do that using the R function *save.image*, For example, if you want to save your analysis in a file named *myworkspace.RData*, you should use:

To load:

N. Matloff, The Art of R Programming. 2011. 1st ed. San Francisco, CA: No Starch Press, Inc., 404 pages.

Adler, J. R. 2009. R in a Nutshell. A Desktop Quick Reference. O’Reilly Media.

A tutorial by two OneMap authors: Introduction to R