Package rafalib. R topics documented: August 29, Version 1.0.0

Transcription

1 Version Package rafalib August 29, 2016 Title Convenience Functions for Routine Data Eploration A series of shortcuts for routine tasks originally developed by Rafael A. Irizarry to facilitate data eploration. Author Rafael A. Irizarry and Michael I. Love Maintainer Rafael A. Irizarry <rafa@jimmy.harvard.edu> Imports RColorBrewer License Artistic-2.0 LazyData yes NeedsCompilation no Repository CRAN Date/Publication :00:40 R topics documented: as.fumeric bartab imagemat imagesort install_bioc largeobj maplot mypar myplclust nullplot peek popsd popvar sboplot shist splitit splot Inde 13 1

2 2 bartab as.fumeric converts to factor and then numeric Converts a vector of characters into factors and then converts these into numeric. as.fumeric(, levels = unique()) levels a character vector the leves to be used in the call to factor Rafael A. Irizarry Eamples group = c("a","a","b","b") plot(seq_along(group),col=as.fumeric(group)) bartab bartab Plot the overlap of three groups with a barplot bartab(, y, z, names, skipnone = FALSE,...) logical y logical z logical names a character vector of length 3 skipnone remove the "none" group... further arguments passed on to barplot

3 imagemat 3 Michael I. Love imagemat image of a matri Produces an image of a matri which matches the natural orientation. imagemat(, col = colorramppalette(c("white", "black"))(9), las = 1, lab = "", ylab = "",...) the matri col the colors las as in par lab -ais title ylab y-ais title... arguments passed to image Michael I. Love Eamples <- matri(c(1,0,0,0,1, 1,1,0,1,1, 1,0,1,0,1, 1,0,0,0,1, 1,0,0,0,1), ncol=5,byrow=true) imagemat()

4 4 install_bioc imagesort image with sorted rows the rows are sorted such that the first column has 2 blocks, the second column has 4 blocks, etc. see eample("imagesort") imagesort(, col = c("white", "black"),...) a matri of 0s and 1s col the colors of 0 and 1... arguments to heatmap Michael I. Love Eamples <- replicate(4,sample(0:1,40,true)) imagesort() install_bioc Install or update Bioconductor and CRAN packages This is function simply a wrapper for bioclite. It first sources the code from the Bioconductor site then calls bioclite. install_bioc(...)... arguments passed on to bioclite

5 largeobj 5 Details Note that once you run this function in a session, you no longer need to call since you can call bioclite directly. Rafael A. Irizarry largeobj What are the largest objects in memory? This function lists all the objects in the global environmnet and lists the n largest. largeobj(n = 5, units = "Mb") n units the number of objects to return units to display, see?object.size Value a named character string of the size of the n largest objects Michael I. Love maplot Bland Altman plot aka MA plot Takes two vectors and y and plots M=y- versus A=(+y)/2. If the vectors a more longer than length n the data is sampled to size n. A smooth curve is added to show trends. maplot(, y, n = 10000, subset = NULL, lab = NULL, ylab = NULL, curve.add = TRUE, curve.col = 2, curve.span = 1/2, curve.lwd = 2, curve.n = 2000,...)

6 6 mypar y n subset lab ylab curve.add curve.col curve.span curve.lwd curve.n a numeric vector a numeric vector a numeric value. If length() is larger than n, the and y are sampled down. inde of the points to be plotted a title for the ais a title for the y ais if TRUE a smooth curve is fit to the data and displayed. The function loess is used to fit the curve. a numeric value that determines the color of the smooth curve is passed on to loess as the span argument the line width for the smooth curve a numeric value that determines the sample size used to fit the curve. This makes fitting the curve faster with large datasets... further arguments passed to plot Rafael A. Irizarry Eamples n < signal <- runif(n,4,15) bias <- (signal/5-2)^2 <- signal + rnorm(n) y <- signal + bias + rnorm(n) maplot(,y) mypar mypar Called without arguments, this function optimizes graphical parameters for the RStudio plot window. bigpar uses big fonts which are good for presentations. mypar(a = 1, b = 1, brewer.n = 8, brewer.name = "Dark2", ce.lab = 1, ce.main = 1.2, ce.ais = 1, mar = c(2.5, 2.5, 1.6, 1.1), mgp = c(1.5, 0.5, 0),...)

7 myplclust 7 a b brewer.n brewer.name ce.lab ce.main ce.ais mar mgp the first entry of the vector passed to mar the second entry of the vector passed to mar parameter n passed to brewer.pal parameters name passed to brewer.pal passed on to par passed on to par passed on to par passed on to par passed on to par... other parameters passed on to par Rafael A. Irizarry Eamples mypar() plot(cars) bigpar() plot(cars) myplclust plclust in colour Modifiction of plclust for plotting hclust objects in *in colour*! myplclust(hclust, labels = hclust$labels, lab.col = rep(1, length(hclust$labels)), hang = 0.1, lab = "", sub = "",...) hclust labels lab.col hang lab sub hclust object a character vector of labels of the leaves of the tree colour for the labels; NA=default device foreground colour as in hclust & plclust title for -ais (defaults to no title) subtitle (defualts to no subtitle)... further arguments passed to plot

8 8 peek Eva KF Chan nullplot nullplot Make an plot with nothing in it nullplot(1 = 0, 2 = 1, y1 = 0, y2 = 1, lab = "", ylab = "",...) 1 lowest -ais value 2 largest -ais value y1 lowest y-ais value y2 largest y-ais value lab -ais title, defaults to no title ylab y-ais title, defaults to no title... further arguments passed on to plot peek peek at the top of a tet file this returns a character vector which shows the top n lines of a file peek(, n = 2) n a filename the number of lines to return Michael I. Love

9 popsd 9 popsd population standard deviation Returns the population variance. Note that sd returns the unbiased sample estimate of the population varaince. It simply multiplies the result of var by (n-1) / n with n the populaton size and takes the square root. popsd(, na.rm = FALSE) na.rm a numeric vector or an R object which is coercible to one by as.vector(, "numeric"). logical. Should missing values be removed? popvar population variance Returns the population variance. Note that var returns the unbiased sample estimate of the population varaince. It simply multiplies the result of var by (n-1) / n with n the populaton size. popvar(,...) a numeric vector, matri or data frame.... further arguments passed along to var

10 10 shist sboplot smart boplot draws points or boes depending on sample size sboplot(,...) a named list of numeric vectors... further arguments passed on to boplot Eamples sboplot(list(a=rnorm(15),b=rnorm(75),c=rnorm(10000))) shist smooth histogram a smooth histogram with unit indicator (we re simply scaling the kernel density estimate). The advantage of this plot is its interpretability since the height of the curve represents the frequency of a interval of size unit around the point in question. Another advantage is that if z is a matri, curves are plotted together. shist(z, unit, bw = "nrd0", n, from, to, plothist = FALSE, add = FALSE, lab, ylab = "Frequency", lim, ylim, main,...) z unit bw n from to plothist the data the unit which determines the y ais scaling and is drawn arguments to density arguments to density arguments to density arguments to density a logical: should an actual histogram be drawn under curve?

11 splitit 11 add a logical: add should the curve be added to eisting plot? lab -ais title, defaults to no title ylab y-ais title, defaults to no title lim range of the -ais ylim range of the y-ais main an overall title for the plot: see title.... arguments to lines Eamples set.seed(1) = rnorm(50) par(mfrow=c(2,1)) hist(, breaks=-5:5) shist(, unit=1, lim=c(-5,5)) splitit split it Creates an list of indees for each unique entry of splitit() a vector Eamples <- c("a","a","b","a","b","c","b","b") splitit()

12 12 splot splot smart plot if n > 10,000, make a random subset of 10,000 and plot. You can also specify a specific subset to plot. If length of subset is larger than n, a random sample is still used to reduce data size. splot(, y, n = 10000, subset = NULL, lab = NULL, ylab = NULL,...) y n subset lab ylab the data the y data the number to subset eplicit subset inde (optional). title for the -ais title for the y-ais... further parameters passed on to plot Eamples <- rnorm(1e5) y <- rnorm(1e5) splot(,y,pch=16,col=rgb(0,0,0,.25))

13 Inde as.fumeric, 2 barplot, 2 bartab, 2 bigpar (mypar), 6 boplot, 10 brewer.pal, 7 hclust, 7 imagemat, 3 imagesort, 4 install_bioc, 4 largeobj, 5 loess, 6 maplot, 5 mypar, 6 myplclust, 7 nullplot, 8 par, 7 peek, 8 plclust, 7 plot, 6, 7 popsd, 9 popvar, 9 sboplot, 10 sd, 9 shist, 10 splitit, 11 splot, 12 title, 11 var, 9 13