R-exercises - Basics of R – Exercises Read the...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Basics of R – Exercises Read the instructions closely! Lines starting with ”>” contain R codes, and they should be written without the ”>” sign. Codes and R outputs are typesetted with Courier font to separate them from normal text. This exercise has been written so that you should test every command, and see what they do yourself. If you need help, just ask! The dataset consists of 17 bioinformatics students, who have given their height and shoe size measurements for teaching purposes. 1. Reading the data into R There is folder data on your desktop. In its subfolder students, there is a file students.txt. Open it in Excel, and check what columns it contains. Note the column headers. You can leave the file open in Excel in case you need to see it later; else just close it. Open R by double-clicking on its icon on the desktop. Go to the menu File, and select option Change Dir. Change the directory to the directory where students.txt file is located. Read the data into an R object named as students (data is in a tab-delimited text file having a title for every colum): > students<-read.table(“students.txt”, header=T, sep=”\t”) Check that R read the file correctly (objects can be printed just by typing their name): > students 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 height shoesize gender population 181 44 male kuopio 160 38 female kuopio 174 42 female kuopio 170 43 male kuopio 172 43 male kuopio 165 39 female kuopio 161 38 female kuopio 167 38 female tampere 164 39 female tampere 166 38 female tampere 162 37 female tampere 158 36 female tampere 175 42 male tampere 181 44 male tampere 180 43 male tampere 177 43 male tampere 173 41 male tampere You can also print the column headers only (sometimes the whole table does not fit on the screen, and this might be more helpful): > names(students) [1] "height" "shoesize" "gender" "population" 2. Prepare the data for analysis Individual columns can be called using the following syntax: first comes the name of the object, followed by a dollor sign, after which comes the name of the column: > students$height 3. Simple statistics What is mean height and shoesize? Type: > mean(students$height) [1] 169.7647 > mean(students$shoesize) [1] 40.47059 What about standard deviations? Type: > sd(students$height) [1] 7.578996 > sd(students$shoesize) [1] 2.695312 What are the gender and sampling site distribution (how many observations are in each groups)? Type: > table(students$gender) gender female 9 male 8 > table(students$population) population kuopio tampere 7 10 Command table can also be used for cross-tabulations: > table(students$gender,studenbts$population) population gender kuopio tampere female 4 5 male 3 5 4. Useful plots Usually graphical inspectation gives an easier interpretation. How are heights distributed? To use a histogram, type: > hist(students$height) That’s the distribution for the whole population. But, is there is a difference in heights between the sampling sites? That can be studied using a box plot. In this case variable height is divided into two groups using the variable gender, and a separate boxplot is produced for both of these plots: > boxplot(students$height~ students$gender) So, there is large difference between the genders in heights. Does the same apply for sampling sites? Write the code for this yourself. How are height and shoe size related? You can get a graphical view of this by making a scatter plot: > plot(students$height, students$shoesize) 5. Recoding variables What if we want to differentiate between males and females in the plot? Let’s use different plotting symbols for males and females. First, we need a vector of plotting symbols. Let’s plot females with F and males with M. The new vector can be produced by the command ifelse: > sym<-ifelse(students$gender==”male”, “M”, “F”) Now plot the image again: > plot(students$height, students$shoesize, pch=sym) Check from the help file what are the arguments for ifelse command. We can even represent different populations with colors. Let’s recode the population variable with color names (Kuopio=blue Tampepre=red): > cols<-ifelse(students$population==”kuopio”, “Blue”, “Red”) Now plot the image again: > plot(students$height, students$shoesize, pch=sym, col=cols) There are only 16 symbols on the plot. Can you figure out where one has vanished? 6. Making a new dataset Make a new dataset from the variables height, shoesize, sym and cols: > students.new<-data.frame(students$height, sym, cols) students$shoesize, Check that the new dataset is OK: > students.new 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 students.height students.shoesize sym cols 181 44 M Blue 160 38 F Blue 174 42 F Blue 170 43 M Blue 172 43 M Blue 165 39 F Blue 161 38 F Blue 167 38 F Red 164 39 F Red 166 38 F Red 162 37 F Red 158 36 F Red 175 42 M Red 181 44 M Red 180 43 M Red 177 43 M Red 173 41 M Red > class(students.new) [1] "data.frame" 7. Extracting a subset from a dataset Make two subsets of the dataset students. Split it in two according to gender. First, check which individuals are males: > which(students$gender==”male”) [1] 1 4 5 13 14 15 16 17 Based on that use subscripts to select the correct subset (take only rows for which gender is male): > students.male<-students[which(students$gender==”male”),] 1 4 5 13 14 15 16 17 height shoesize gender population 181 44 male kuopio 170 43 male kuopio 172 43 male kuopio 175 42 male tampere 181 44 male tampere 180 43 male tampere 177 43 male tampere 173 41 male tampere Similarly, make a new dataset from females. Sometimes we want to split the dataset using some continuos variable, such as height. Typically the median of the variable is used. Make two new datasets that containg individuals below and above the median height: > median(students$height) [1] 170 > students.short<- students[which(students$height<= median(students$height)),] > students.short 2 4 6 7 8 9 10 11 12 height shoesize gender population 160 38 female kuopio 170 43 male kuopio 165 39 female kuopio 161 38 female kuopio 167 38 female tampere 164 39 female tampere 166 38 female tampere 162 37 female tampere 158 36 female tampere Similarly, make a new dataset from long students. 8. Quit R To quit R, type: > q( ) R then asks you whether you would like to save the workspace or not. This is generally a good idea, and answer the question “yes”. You can then get back to the same analysis just by double-clicking on the .Rdata-icon in your students-folder. If double-clicking does not works, you can start R, and use menu choise File->Load Workspace and File->Load History to acquire the same result. AFFYMETRIX PREPROCESSING EXERCISE PRELIMINARY OPERATIONS A. Start RGui; B. Change the working directory shoosing the folder where thr CEL files and the PHENODATA are located; C. Load the needed libraries: library(affy) library(affyQCReport) library(hs133ahsentrezgcdf) library(hgu133aEG1000) In this exercise we will perform the following tasks: 1. Importing the CEL files and PHENODATA; 2. Performing basic QC; 3. Preprocessing the data. 1. IMPORTING THE CEL FILES AND PHENODATA First, we create a new AffyBatch object (dat) where we import the CEL files. dat <- ReadAffy() The PHENODATA contains the information about the samples (i.e. the microarrays) of our dataset. The PHENODATA are usually stored into a table (TAB-delimited .txt file) where each row represents a sample and each column represents a variable. We create a data.frame object to import the PHENODATA text file pd <- read.table("phenod.txt", header=T, row.names=1, sep="\t") Then, we assign the pd data.frame as the PHENODATA of the dat AffyBatch pData(dat) <- pd 2. PERFORMING BASIC QC We use the affyQCReport library that we loaded before QCReport(dat) This creates a PDF file with some plots. Read the affyQCReport vignette for interpreting the different plots. Note that the plots in page 2 can be also produced by the commands boxplot(dat) and hist(dat) Additionally, we can check the RNA quality by using the AffyRNAdeg function. deg <- AffyRNAdeg(dat) Individual probes in a probeset are ordered by location relative to the 5! end of the targeted RNA molecule. Since RNA degradation typically starts from the 5! end of the molecule, we would expect probe intensities to be systematically lowered at that end of a probeset when compared to the 3! end. On each chip, probe intensities are averaged by location in probeset, with the average taken over probesets. We can plot the results into a new PDF file: a) we create a new PDF graphic device. this will direct everything we plot into the new PDF file pdf("rnadeg.pdf") b) we plot the RNAdeg results plotAffyRNAdeg(deg, cols=1:17) c) we add a legend to the plot legend(1,60,legend=rownames(pd), text.col=1:17) d) we close the graphic device dev.off() 3. PROPROCESSING THE DATA We want to use the re-annotation of the affymetrix probes of the hgu133a chipset according to the Entrez Gene database. For doing so, we instruct R to use the new CDF package with the re-annotated information dat@cdfName <- "hs133ahsentrezgcdf" We also instruct R to use the meta-annotation package dat@annotation <- "hgu133aEG800" We preprocess the data using the RMA algorithm. The new datrma object is of class ExpressionSet datrma <- rma(dat) Finally, we save the normalized expression values into a TAB-delimited text file (.txt) write.exprs(datrma, "datexprs.txt", sep="\t") FINAL OPERATIONS A. Save the workspace save.image("Affy_Preprocessing.Rdata") B. Save the history savehistory("Affy_Preprocessing.Rhistory") Dario Greco Institute of Biotechnology - University of Helsinki Building Cultivator II, room 223b P.O.Box 56 Viikinkaari 4 FIN-00014 Finland Office: +358 9 191 58951 Fax: +358 9 191 58952 Mobile: +358 44 023 5780 Email: dario.greco@helsinki.fi o 20!!0 ‡©0C ¥ ˆC 2  ‡ ¥ ‡ ‘ 2   —C  — CC  C E0 ‡ ˆ — d B0ˆ ¥ ‡ ‘ —  Î 2 $   2 •fF¤Íf#€„t•…f¨©('Èl4¨CQ¤v04%”•c'¨C(u¤Ï͇ ¨©('q„—C f’ͤ0 x ’$ ˆ  B — § $ o Ì Ë Ê “ ‘ § d ‘   0 " C d 2 ¥ ‘ "‚ C Ç ˆ — d B 0ˆ   ! C ¥  C ! ‡   0 ˆ —  0 0  u(„¦`W”¢’7Qvɤ@7Q†Ä¨„—'XFvƒÈ¨C(u¤fTAC QÆÅ–C (FĤ“¨C‰f¤€3 s C ¤†C ‚p !0 !C‡ » ÂÁc7A 7™ °– ¡ š ¤ ½ ¾ ½ ¾ ‰µ À £ ¢¡ · ¶ ¢ ÂÁc7A 7™ °– ¡ š ¤ ¼ ¾ ¼ ¾ ³ µ À £ ¢¡ · ¶ ¢ º ÂÁc7A 7™ °– ¡ š ¤ » ¾ » ¾ #µ À £ ¢¡ · ¶ ¢ ÂÁc7A 7™ °– ¡ š ¤ ³ ¾ ³ ¾ ² µ À £ ¢¡ · ¶ ¢ ¹ ÂÁc7A 7™ °– ¡ š ¤ º ¾ º ¾ ‰µ À £ ¢¡ · ¶ ¢ µ ÂÁc7A 7™ °– ¡ š ¤ ² ¾ ² ¾ lµ À £ ¢¡ · ¶ ¢ à ÂÁc7A 7™ °– ¡ š ¤ ¹ ¾ ¹ ¾ zµ À £ ¢¡ · ¶ ¢ ÂÁc7A 7™ °– ¡ š ¤ µ ¾ µ ¾ ¿ À £ ¢¡ · ¶ ¢ ™A‹ – 77a°7™ °– ¡ š ¤ ´ ¤´ ½ ¡ hj¸ · ¶ ¢ ½ ½ ™A‹ – 77a°7™ °– ¡ š ¤ ´ ¤´ ¼ ¡ hj¸ · ¶ ¢ ¼ ¼ ™A‹ – 77a°7™ °– ¡ š ¤ 7´ ´ » ¡ hj¸ · ¶ ¢ » » ™A‹ – 77a°7™ °– ¡ š ¤ ³ ´ ³ ´ ³ ¡ hj¸ · ¶ ¢ ™A‹ – 77a°7™ °– ¡ š ¤ ¤´ ‰´ º ¡ hj¸ · ¶ ¢ º º ™A‹ – 77a°7™ °– ¡ š ¤ ² ´ ² ´ ² ¡ hj¸ · ¶ ¢ ™A‹ – 77a°7™ °– ¡ š ¤ 7´ ´ ¹ ¡ hj¸ · ¶ ¢ ¹ ¹ ™A‹ – 77a°7™ °– ¡ š ¤ ´ ´ µ ¡ hj¸ · ¶ ¢ µ µ ³‹ •± ²‹± ™km¬™hj¯ ™km¬ Š™ˆk­¬™j •—c‰7l7°¤®(7A‰h « š ™˜Š – A7m ½¾ ¼¾ »¾ ³¾ º¾ ²¾ ¹¾ µ¾ ½ ¤´ ¼ ¤´ » ´ ³´ º ‰´ ²´ ¹ ´ µ ´ – —‡ rˆ ¥0 § š ™ ˜ Š —  ¥!0E C ‡ 0 ‡ ‘ n§ — ©C #„¤fl© – A7m – ¨C(¤v0F¤7eFt¤c'–A¨§vªˆ#¨§  – ¦ –¤ š ™ ˜ Š m ž £ ¢ ¡ Ÿ š ™ ˜ Š m ž  m ™ › š ™ ˜ Š – •Ÿ lA¥S– A7za(l‰Œ c– (•7AŠ œ– A7m —‡ r  Cˆ 0 0  C ‡   §  0  ¥ 0 B! §‚ ¥ ˆ ¥ 0  § © ¥ ‡ #ff¤WFt(%¤1¨§F¤gFƒ„v„¤fl˜‡ ¨©('‘  ‰ o C …!0 p“ ‘ § d ‘ —  C ‡  0 !  0  0© — ¥ 00  C ‡ !C ‡ ‘ 2! © ¥  … ¥ n! § ‘ C ‡ C …  0 ‡© C ‡ z• #ˆ •F¤X”¢7%’v¨CSFW#C w¤TC ¤#¨§Ž¨ˆ‚f¤xF`C 'WF§ C „!¢(„AF¦`Fx•(¤cf”‚p  m kkj hŒ ‹Šm Šˆj (cl#f`l7l‰h ‡ … ¥ d 2 ! §   C B B !‚ ¥ … ¥ © — ˆ 2  ˆ   ¥ C … 0 n© 0 d m k k j h C n 0  C ‡   0 ! 0ˆ "  ¥ („(†v…lD„§ ƒ‚„D(„F¨CC 4AeC u§ l„”•¤fc“cl(T•¤f€F4¤vF¤fF(#Fˆ „!@~  H U R sb u  u | a'†}fta'ph t{  ¥0 —0 B!§  ¨§F¤z„y„¥¤gF'o x ˆ ¨§—(w¨©F 77¤v0t5 d ¥ ˆ§… ¥ o !§ d ¥ 00  F%u„Buf¤to s r! 7C ¤0 ˆ ¥ d§ p oC …0 ©0 d m k kj h ‡ ¥ ‘ 00  ˆC© ! d ! d  ‘§ ‡ !0 —  … ¥§ … !0 C ‘ " ¥ !C ‰C ¥ ‡ … ¥!  ¨©(qv7•¤7n cgcl(iF„'gf¤eˆ § (C (D§ ™#˜F¤#¨CD§ –(„•”C ¤“’%7C ¨ˆ© (¢¨ˆ(F†(„F(¢ƒ uu y wuHHuUH b a‚f€xQa'v%XV ts §  E!C B§ &  0 FD1A¤@9 q i hV e bYYPV UH RPH rp@gfd ca`XWQTSQIG 8665"3 20 )" $& $ "! ©§© ¥ £¡ 77741¢('%#§ ¨§¦¤¢  ¶ ˆ˜Š ˜¢hŒ ™ ¢h  l Dµ ¦ H7¤ lA•Œ ¹ •lf`‹ – j š ¥ Œ – A¡ ‡  0 B B§© … ¥ˆ  2  C ¥0 § ¥ B0!…  ¥ ‡ ‚§  ¥0  ˆ ! d ! !C ‡§ § B ¤gug(„F–l…„¤f(t¨ˆ„¤f7§ F¨ˆ(S¤1¨§F¤flC C (C tFDA&  ¶ ˆ˜Š ˜¢h – ©  A Dµ ¦ C¤ lAFŒ ¹ •lό ‰š j B‡ F B! ‚ˆ 0! ¥ B ‡ ¥!0 … — … ¥ — d d0 ! ‚0 C … 0 ‡©  ¥  ¥! ¥  C ‡ ‚§ C d0 ‡ˆ C ‡ ˆC § ‘§ GF§ •¤fF”¨©u(F„F¤l¨§D(„A„2((¤vC ƒ‰T•(¤cfͨ§F((„FFF¨ˆWFS¤wQ¤cFWF'( #¢E  ¶ ˆ˜Š –  Dµ ¦ C¤ lAFŒ ‰š j © B‡ r s fF¤0 § ƒ¨(¤cf€C €„'„¤7Ž¤c7„…F…C fF(( …(§ 7n cS¤Í¨§F((„FFF¨ˆWF–A¨§vªˆ#¨§ 2 0!! B !‚ — C   0 ‡©  ! ¥ ˆ C — 0 E — 0  ¥ˆ   © 0!  ˆ    ! … © 0  ‚ §  ¥  ¥! ¥  C ‡ n § — ©  C Y  @ R Vb  Y U d e ub | a‚U Apfa(tfttfph w9 Ÿ·h 7%•lm © Ÿ › 7¢ – cg8AFŒ – ´ AAAl˜ ¾ ´ ˆ  © ™k ¦ ˜Š ™ŠŠ¢±­¢Š m 64 ˆ˜ 75lAŠ ‡ rÎ w„¤7gA4v0(C eC ¤“Î 6 #¨§Q'„¤Q„—¤c¨¨§FfF(( €(§ 7Afc'C  ‰wC 'ÈFDeC ‡ o 6 C —0E 2  C© — d ! !0 o ‘ —C  ˆC —0E —0 "  ¥©0!  ˆ    !… n©0  ! ‚0 !C ‡ ‘  § ‡C B !C C ’ªˆ#¨CQ#•„EF¤cFC „¤`¤f•C È(Dv’o – ´ AAAl˜ ¾ ´ ˜¤guÈ(„Ff(’¨ˆ¢¨ˆ(FgAl#h ˆ  © — "ˆC ¥0 ! —0 —0!C E ˆ … § B ‘ ™ Š Š¢ ±   ­¢ Š m ˆ  0 B B§© … ¥ˆ  C §  ¥ ¥ ‡ m k k kj  W•†„FC A„…(§ 7C § wF}§ ƒ…C fF(( ’¨ˆ’%Fif#’¤Í2 „FC l„˜(§ 7n c@¤#A¨§4„2„—¤cF¢3 “ o 2 ¥ˆ   ¥    ! … ! ‚ C ‡ B !‚   © 0!  ˆ ¥  § dˆ ‡© 0 C ‚ § ¥ˆ   ¥    ! … © 0  — 0 © — — 0 ˆ d bVPHHU y d 1UHe 0V vU aq`%%'@V fT2%'W¨tu (| )e' à %  ™™ ˜ ¦ l”¹&µ §š ¥˜ ¤ AŠ ‡ r 20!!0 C ‡ § ˆ  ¥0© — ¥C ‡  0 ˆC B0  § dˆ 65  !  C ‡ CC %#fF¤”F€t¨§F¤#¨§t„!F4¤¢D¤c'%F“7¢Fˆ ¢FW&  š $˜ ¤ A•Œ – 7# – ˜ › ™ – !al¡ ¤ AŠ ‡ ™™ ˜Š ­¢‹m ™ Š jŠ ˜ Ÿhm £ ¢¡ ŸŒ # "m™ › ™™ ˜ •cl˜ ¤ c(lz@•7l(Š Áš ¥˜ ¤ AŠ ‡ 2   Cˆ   0  0 ! ! 0 ˆ C ‡ p o  ¥ 0 B! §‚ ¥   § 2 — 2 0!! 0   0 ˆ C B 0  C  C … ¥ 0  § ©  §  §  ˆ —   § d l'f€¤…#C eC ¤”C ‚’¨¨§F¤gFƒ„’(•v0ÈfF¤€¤¢D¤cW•…„¤fAv“(v¨C‚QA& à › !µ A› › Ÿ – ¡ š 7(Ÿ ¤ ™ – ¢ ¤ Š l© ‰((¸ ¤ AŠ ‡ ¢ mˆ Š © ˜  š– ˜j™ ˜  ¶ ˜ ŠŒ h ¦ ¶ ˜ ŠŒ ¸ ¢ Š  l ¤ A•Xl¢ ´  7¤ AF`l¦ µ Œ ¦ aŠ – k œl© ‰((¸ ¤ AŠ ‡ j m › š– ˜j™ ˜ „o#QF(c © u(„© ©§ dˆ 0  B —§ 00  f¤…„¥ s ¨ˆW„¤7WC 7''ˆ QFWC 7F¨„—¤†§ “¢A7„…’W•„E7”’g(„'z¨§„—ƒWF4†xzˆ %F…c@¤vŽC ¥ C — 0 E ˆ § ‡ ‘  § dˆ ˆ § ‡ — 0  6  ‡ ¥ C ‘ C ¥ … C ‘ … ¥ ‘ — §‚ C ‡ “ o  § dˆ  0  ‚ §  © C Å C ‡ !§  ¥ o o 0© C ‘ ¥ ‡ 2 £ o § dˆ C ‡   ‡ ¥C ‘   !C ¥  C ¥…  — ¥ ˆ§ d ¥ ¥ m k k kj FgC 7„…†… 7C ¤#g¦w¨ˆ(F1¦†#ˆ QF†F4§ fˆ l7„…’flC ŽÆÅg•„E7€§ D¨C(„Fˆ 7QW¨ˆT„4cll#h  “ o j¡™™˜ 7Ÿ ¦ a¥#Ÿ › ™ ´ A¢ š f€¨(’w’‘ Š­ Cˆ  — § ‘ C ̈ —  ! d oÇ 00  ¥ 9C C ¢ !§ ~ o ¢ ϨCx(7… •uf¤uƉxeS7Ÿ – ¡ š Ÿ › ™ ´ l7¢ š cƏQFt•ct’7#¤gF§ `QA†„S¨ˆef¤x(t„A& Š ­ C ¥©C dˆ C E0 ‡ C ‘"0 B! ‚ § d & ¥ ¥ 00  ! § C© ¥  ¢ mˆ › k­h Š © ¦ ¢ 7Ÿ – (¡ š •Ÿ ¤š •l¢ ´ ¤ ™ – ¢ ¨Ÿ – ¡ š Ÿ › ™ ´ l¢ š §‰(A¯ ¥c– m – Œ š •(k ¤ 7AŠ › AŠ ‡ Š­ ¦ ™km¬™hj ¤š ™˜Š ™˜mkjm m™ ˜ „r#QF(c © u(„Df¨ˆ¤Í¨§F„(̈́4¤'¨ˆ(¤c‡ …C ¢£¤€v ©§ dˆ 0  B —§© § —0  ¥ ¥  0 ¥  0 Ì —C  0 © C ! ¢  0 C B !‚ ˆ C — 0    ! … n© 0 ! § Ç ˆ  B — § © —  ‚ C  C ‡  0 ! C ‘ " … ¥ ‘ — §‚ C ‡ “ ™ ˜ m k j m  m ™ § ƒw„¤7E (§ 7Afc ¡ C ƒ‚ 4u(„“„(¤0 (†F #C g’`(„'z¨§„—ƒDF  4o š 7c(k ¤ 7AŠ  0 B B§© 2 ˜ Š —  ¥!0E   ! ˆ  0 0 ! ! B0!… ! d ¥ˆ —0 0 C …0 ¥ C ‡ B ‚ ˆ —  00 ¤guÁA €¨C(¤v0F¤7i§ œC § F—¤°#C €C ¤0 ¤f7§ (D¨ˆF„2¤c¤Í•¤g„B€F § ! g¨C˜f¤Wƒ  l ¤ ¹ ”kŒ š c‰($l¢ ´ š ¤¦  "¤ ¹ k Xl¢ ´ ("¤ ¹ k Œ – A¡ ¦ ˆ ‡ m ™kmh › ™km ¤ m Œ h &¤ m ¢h ¢ ¤m £¤ k Œ š c‰¥l¢ ´ ¤š ¤ "¤ ”qA¢ ´ '£¤ k Œ – A¡ ¦ ˆ ‡ ™ k m  h › ™ k m ¦ ¤ m kŒ h & ¤ m ¢ h ¢ ¤ £  l r©C ÅC C ‡ CC #ŽWFW& Ÿ™h 7allm ´ š Ÿ › •¢ – †¦ ck Œ š (llA$A(¸ – §!(lc7( › ¹ k ‡  © ™k m ‹mŠŠ ™™ ™#™jhmkŠ¢ m … ¥ˆ  C —0 B!§  C  0© ˆC©  ¥! („F…#„y„¥¤gFe%…¤#v¤v0F¤0 %  Ÿ ¢ m j h m kŠ¢ Š · j m k ‹m Š Šm¦ m š ™ ¢ # ¡j j ¢ h •Q¥j – £lc77¬ ™ – AŸ ›  c¥¦ µ › 7l7F§k Œ lš lzaž –  cŠ ¡– A¡ ‡ 7Q¥j – £A77œ77($Ÿ › !(†¦ µ › (lm ¦ à k Œ lš lzaž –  cŠ ¡– A¡ ‡ Ÿ ¢ mjhmkŠ¢¬ ™Š¢·™# jmk ‹mŠŠ m š ™¢#¡j j ¢h  Ÿ ™  ¢  © ™ k¦ ˆ ˜ Š ‹ m Š Š © j  ™  j h m k Š ¢ •c¥ Ÿ › •¢ – †8lA•Œ š 7lAl j – a"!(lc7( › à k ‡ m rf¤0Í#y„„¥—¤0gF!§4¤01#y„„¥—¤0gF!((4F%'§ ‚ Aš l‰cž – !aŠ ¡– l(4F¨p 0 C B C B §    ‡§  ! š ™ ¢ # ¡j j ¢ h ¡ 2! » A ‰µ ¦ ¼ Œ ´ › (h ¦ ¦ µ › Am ¦ ”S – A¡ ‡ kj ‹ m Š Š m kŒ ¤ ¢ h r — ‘ #¨§(d ˜) 0 C ˆ C 0©  ¥0 —0 B!§  ‚§ — ˆC oŸ ™ ¢  © ™ k ‡ ¥ C ¥0 § ¥  ¥0 —0 B!§  2  I„¥ C DQ ¤#Á¨§F¤z„y„¥¤gFt¤”„( v47a¥ Ÿ › •¢ – ÄF„'‘ „¤f(f¨ˆI¨§F¤z„y„¥¤gFil¤0  § ‡ ¥ ‘ 00  ¥0! …§ § o š ™ ¢ h ¡ – j Š ¡Ÿ ! ‚ ˆ  0ˆ Ÿ ¡Ÿ o oC " ‡… § C ¥  § ‡C B C ‡ ‚§ !  (F„'–f¤…¨§F¤f4¨T•Ÿ Aš lA‰j l– !al†§ e(¤fFuQ€… 7X¨7(w¨ˆ1AFD“F¦¤”C F¨C—  !  C ‡ — § … ¥ ¥!  o š ™ ¢ h ¡ – j Š ¡Ÿ !§ " š ™ ¢ h " Ÿ mj  ™ kŸ o oC C  0©  © ™ k  C B …! Fˆ TF̈́2(€(„F„Fi7Ÿ lš Al(zj A–  cl†Q7Ÿ Aš AlzŸ •Q¥•–… 7“%1¤#€•¢ – WlD(7F¤0 Ÿ¡  © ™k ˆ˜Š ‹mŠŠ  © j™jhmkŠ¢ 7ŽŸ › 7¢ – c¥¦ •AFŒ š AlA!j – %££A77 › k ‡ m r 00   ©0!  ˆ    !… ©0 f¤…C fF(( …(§ 7n c C ‡ B !‚ ˆ ¥ 0! … —   d B § © ©  — " — C   0 F€§ ƒf7¨§F¤fi¨§4C ((uuªˆ#¨C¦#¨ˆ(¤c‡ © x ’—¤1’w¤D7¨§F¤fÁ„ÁC FC C A„ÈC ¤€’È„A& 2 $  0 Î 2 $ ‚§ ˆ ¥0! ¥  ˆ !  ¥ !0 C ‘ C© ¥ d bu u sH vU a‚@b @ tŽ'U i | ¹ A ¤µ ¦ ¼ Œ ´ › (¨Ž$l(•#Ÿ A© ‰j kjh ¦ Ÿ ™™Š˜ › ˜ ¹ A ‰µ ¦ ¼ Œ ´ › (¨%•Ÿ A© ‰j kjh ¦ Ÿ ™Š › ˜ C ‡ p oÌ  ¢¨ ‚fa¢¢ © ¦ c™ – j © Ÿ › ¤AÍ8™ –  cl¡ ¤ – Ÿ ¸ ¸¢h ¦ Š jŠ ˜Š ¦ © ¦ c™ – j © Ÿ › ¤AÍ8™ –  cl¡ ¤ – Ÿ ¸ ¸¢h ¦ Š jŠ ˜Š ¦ Dµ ¦  •" ¤ AFŒ ˆ ˜Š Dµ ¦  ˆ ¤ AFŒ ¶ ˜Š  ¹ •lό – A77cj ‡ ˜¢h ¢h¡™˜mk ¹ •lό – A77cj ‡ ˜¢h ¢h¡™˜mk rÌ …! § 0 ˆ ¥ a7FS‚¤€#C F„D…F! zeC ˆ ÇIC ¤l„…¥FFC•l„¥eQ4¤#©€†„FˆC A„€(§ 7n cS¤q•¨C— ‘§  0 ˆ E C 0 2 ¥  ¥  !… ©0  ‚§ —C E ¤0āt†© ƒœAFDBi¨§F0fF(( ˆ (§ !7n ©ce§%d S¤0#¨§—“¨C(du„BFˆ1C uC¦Ž•„2F¨§C `9 ¨ Ç § ‡ C  ¥ © !   … 0  ˆ — © — ¥  ˆ ‘ " —ˆ  ¥ E ! š ¦™  ¹ Œ – ™ F¤ § ‡ r7C „¥F'("¨„¥¤0l…¤0'•5 Ç ‘z(„'‘¨¤0c„7…„F§wF€„'¨§(D§ vl¤‚€•X†“ !‘ Ì § ¥ — ¥ ¥! C ‡ ¥  — d   0 ‘ § 2 ‚ 7Ÿ – ¢l(¡ ¤¥š j š j © Ÿ ›  m†¦˜7Ah¤6 ¶ 7¢AhF¦ ¹ ¢£ ¶ ˜7¢Ah ¡"˜7¢AfŒAŒ – A¡ ‡ h ž jk "¢ ˜ h ¢h µ¦ Œ Aµ —‡ rw#§(„¥'‘ Cf0˜%da•„2F¤f„(”„BF’ˆ ¨§(u’‘ eC D§ t( € “ ‘ ‘  C § " —ˆ §C 0 —  ¥ˆ  — d §  C ˆ  !C !§ – 7Ÿ – ¢l(¡ Š7™ A– m ´ š®š j š j © Ÿ › !(†¦ ¶ •lh •lό – A¡ ‡ h ž jmk ˜¢ ¦ "˜¢h ¢h  ¶ ˆ˜Š ˜¢  !µ ¦ C¤ •AFŒ ¹ 7lh › ¶ •lh ‡ ˜¢  !µ ¦  " ¤ •AFŒ ¹ 7lh › •lh ‡ ˆ˜Š ˜¢ "˜¢ o  — d ‘ )   0 Ì Î 2 $Ç … — ˆ E 2 $Ç … — ‚ §  — d !  0 ˆ C ‡ ! 0 d B §© © C #¨§(ͅ¤'•’¥c¨§‚¦Ì x ’¥c¨§@¤’¨§(tC F¤#© wFWC ¤cu’ªˆ#¨§ ¯ ¸ º ¶ Å ³ f ² ¬ ­ g ¬ ¬ ¹ Æ Å e Ý ³ Ü e ™ g f ¬ » ˆa¸ µ ˆ» •° bAF® ´ A•A~e A° %¤¤A77Æ µ f ª g …‘    ƒtC '¨¥  § € C B  ¥ © !! § ©    ! x h© 0 z ‰  0m C   ¥ 0 ‰ 0 B! §   C ‰ ¥x ¥! — € ¥ … 0 0  C € C ‰ 0 B! §  Ø  C FD}¨§FC FR(§ 7A„tŽw(¤ƒ(o4¨§F¤~wÛw¥¤rFa ¨§}w—FyÏAwF(‡Fw'k„¤ŠFb#wÛw¥¤rFÚف#Ž‰ ¯ ´ ² ¿ AÏ® ° 7× ° ² µ ³ ° ~|•º ¿ A¬ ª ³Æ³ ²¬ û­g ³ ¬ Æe¬ ² ¯¸™g ¼¹»º¹¸® ×¹Ö¶g³ µ ³Æ³ ² ˆt•² ½ t(•~@ˆA7•(¬ Õ´ 7² ¿ A¬ ª …  ¥ 0 B! §m ¥   § 2 ‰ 2 0!! 0   0  C B 0  C  C x  0 ! ‰ 0 ¥ Ž¨§F¤rFwd(ˆ–0{#„F¤¥¤nD¤trˆ4#C {w2w‰¤tw@Ô Ò Ñ µ » ¶g« ¬ Å ² Ð ´° ²e Ë ² Óµ ~Ì Aµ ¸ ° º ´ 7(¸ ¿ ³ •° » ¿ ¬ b•Å €(³ ¤¿ A¬ ª ¯ Î ² ¬® ™ » Ä Æ Π² ¬ » ¬ ÆÂ Ì e g µ ´ ° ² e Ë ² •¯ ¿ Aό•7t£Ž¯ 7¿ AF® Ë •Ía~® ¾ |¬ ° f •Å €(³ ¤¿ A¬ ª …6 € ¥C … BC€ ¢•7wx¦ÊF x ¥ ¥ x 2 z ‘ p C ‰ 0 E  0 € È  § —  0 z È  B ‰ § ©  0 z  C h! 0 B  C C z C E 0 € ¥ € …  §  C €  0 ¥ ‰ C  ‚ C (wwE7y•b~‡w¤†t'~aF(tƒÉy(wodtrˆF¤rragˆt˜€ ¨©('ǁ %— dFnC ¤twywBw¥Çƒ¢s ¯ » ¶g« µ ÆfÙ» ¬ Å Â » 7¸ ° (º ´ ˆ¸ ¤´ ˆ•7Ä ½ ³ A° » %¸ ° º ´ ¸ µ |•» ´ `€(Ae ˆ7t° g ° ® ´ ˆ(f ½ 7A¬ µ A¬ ª ³Ä¬Ã  ³fgÁ³™ À¿´ ³²¬ ³²gfeg ¶g³ ² … 00  C € 0 ! C € „¤}F4#C „kj ¯ ° ¾ °½ ´ ³ ² ¬ g · ¼ ¹ » º ¹¸ ´ ³ ² ¬ g · ¶ g ³ µ ´ ³ ² ¬ ° ˆ¸ •Ay`° A7~|(•€® t° (ˆ7A¬ ° A7g ±ª … ‰ ‘ C x!0 C € 0 !  0 2! © ¥  x ¥ h!§ … C € C x 0 € 7¨C' ˆF¤„‹F{#C 4¤„F§ C w!˜(wAF¦‹F‹ˆ(¤t†$ ¯ g ffe ™® ­¬g ¬«e (t•#„b•7•€™ ª C y§ Aw‡ˆ¤7h ty•#Rˆ¤„{¤udF¤„A&    ¥ C x0 ©0 — g f fe ™ C h0  0 !0 s HYe H © fW § bW a'SP v„H baY ¨¦ i‘ 2! © ¥  0 0  ¥ m  i 6 6 6 5 … ‰ 0 $ ¤ 5 6 5x 5 5 6 5 ¢ ¡ Ÿž —œ— › — š ™ ˜— ƒ(F§ C w!y„¤„w†%— 777d~¨§w‰¤ky¥777£€777" ’ƒFFƒgrD7|ƒd– i7¨©yS•¨Ctƒ(rn“w“¨§F C (ƒ‚C C w¤”Fw'‡ˆSt“‘F•(¨d(ww¥¤§ (“¨§F C (ƒ’„F¤ˆ§ ¨©d) C ¥ B   ¥© ‘C  ‰ v • ¥  ¥  ! — ‚C  ! ‰0 € ¥ … C C x C ¥ C ¥ x ‰ ‘ ! —  ¥  ! — ‚C 20!!0 ! ¥ iƒˆ777'ƒŠg†Žw((g¤kw"Œb‹¦”aAkw"Šbd†(rkw"%g#w(n†w"bƒ¦‡~¨§w‰¤kd7¨C¨©FF¤„F ‘ 6 6 6 5 i ) i ˆ " ¥ z    0 i 9 i j "  C C — & i ‰ i ˆ " x  § ‡ i & " ¥ §   v i † i ) " … ‰ 0 $ … ‰ ¥! 0 C € w‚(oraƒ¤#DADwBF%ƒ˜Fd¤y¨w‰¤„rI~ˆSF¢¤‚wwEF4FS a€4ˆr¢7%nn¤§ ¥    §m C z  0 ©   C ¥! C — ‚ C C € m § ¥ 0 C v i  C  C x ! C € § m § 2 ¥ ¥ © 0 C €  © C ’ 0 C  C x w t § — t m (ˆ„AnF~†({}§ ¦¨|¤ˆˆ¢k•S¢drC oƒC ˜Fw'{C ¤ty{%¢ˆt¢¨©ya¤rF  § h© §  h C € … § €   §   ‘  ¥ ‰ 0 § x C € j i t s C ©  ! Cm ! € ¥ …  ! 0 — B § ©  C C z C E 0 € C ¥ B ‰ 0 B! § 81¤S¨©y¢(ˆh yy¢7%nXv§ ounnRFSC 'r„F¤ˆ§ ¨©yr#q§ on¨D„¤lkr7ˆ¤7h t—   0 C ¥ B   § © §  hxw t § — t 8 B !m t s C € ! C € …  2 0!! 0 ! ¥ B 3 p B !m ¥ 0 0  C € j i C x 0 © 0 g f fe ™ € ¥ … C C x   ! — ‚C ‰ ¥  !C ¥    ‘  …§ € !0 ‰  x ¥§ x !0 C … "C ¥ !C ‚C ¥ € x ¥!  t•#dFw''ˆ˜C  C (ƒ4w2w‰¤–0F•C ”“’4R§ #4F¤#¨CS§ y(wˆrC ¤‡†7„¨© ƒ'¨(Fy(wF(¢v sW PWT QWssWY q pW h f HcPWYW H Q T H Q H bRbgubttbaSr@iFf ge %dba`XV RUSP SRP IG §  E!C B§ &  0 FD1A¤@9 8665"3 20 )" $& $ "! ©§© ¥ £¡ 77741¢('%#§ ¨§¦¤¢ i 0.639 ¨‡¨©y‹%„¨w‰'{¤}¨©yd(ˆ„A¥¦arDˆ@¤r¨§F¤„7¨§‹Fl§  ¥ C ¥ B C — 2 ¥ …   0 C ¥ B   § h© §  h CC …C z p C C x m § ¥ 0! x ‰ C € TGPtrqGS ŠÒ |rrrrqS ÇÒ  VÌEVE ½ Ì$EFVF ½ ¶™eÝ! 77|##9  ¶™e G³ Î p77|Ý Ñ  ³ » ¿ e Í”Ì Ð G•7Ä °  ª … ! 0 C  C x ! ‘ C € ! §m  C ‰ 0 E !  C B 0! 0 — C € j i C  C 7C ¤}ˆ'  rF'o'w¤nC D¤„¤t‹kŠ7ˆx €© 0 C ! §m   0 ¥ C !  C B 0! 0 — § … C z ¥ … ! C € "   B ‰ § © § …  C ¥ 0  § © ¥! 0 B  ¥ C  C © ¥ „#DoÕC ¤rwBFFr C D¤„¤tɦ{agw‰w‰'{C F¦#y(w’¦w¤„A“‚FF¤rÕ7wxF(„wA& ¯ Æ ² ³ ¶Â g f |•€e ´ Ï@Í® ° e À ¤™ µ °  ª f e … x ¥ ‘ ‰ C  § B ! 0 C ‰ x   0 0  C € B !m   0 ¥ C  C € ! 0 !  C B 0! 0 — C € (wFF”(Dg¤#ww¥D(w¥ l„¤‹Fʧ o{C ¤rwBFF¥F‹C ¤¢ C D¤„¤t‹kj m C 2 0 c ¥ ¥ 2 0 ` m C x ¥ ¥ c ‘ ¥ ¥x 2 0 ` ‘ m C x ¥ ¥ ƒdÇx 3gi¨w‰hx 3gfƒdÇ{¨w‰edt¨w‰4d3bˆp aƒƒd¨w‰5ˆp …C x C …  20!!0  € ¥C  ‰ C € B ! #ˆ}†'„F¤r•7wxgF–0‡Fʧ (Ô i ©  ! Cm   C —  ¥  C … C z ¥ 0! x ‰ C €  C x C …  2 0!! 0  € ¥ C  ! ‘ C € B m  0 €   0 C B ¥ € 7CC ƒC !y¤0da2y¨w‰'…˜C†£%}¨§F¤„7¨§'FŠˆg¦Š„F¤k•7wxÇF 'F § ! ǤtFŠ¤#DX€ ¨©('… m C x ¥ ¥ c ‘ ¥ ¥x 2 0 ` ‘ m C x ¥ ¥ ƒdÇ{¨w‰edt¨w‰4d3b(6 aƒƒd¨w‰5ˆp …  © ! ‘ C € j i 2 0!! 0 © 0 B !m ¥ 0 z §  0 © C …  ¥ 0 B! §m ¥  0 € … ‰ #¨¥rC 0#gF ‹kLj„F¤I€ #C § o4w¤„(¥¤#‡¦„¨§F¤rFwg¤t'Žw‰C  È ¥ 2 0   Èm Ž¨w‰¥4xd3)Ȓ¤0rFƒCnx¨w‰¥ RyB(w‰©D†£4wS~‹w¤rkŠˆ#„F¤‡y§ '(aFC F#§ „„Lj ¥È  § § … ¥ 6 ¡p C ‰ 0 E C € j i 20!! 0 C  §     § — !! §© … ! €© 0  G³ Î ° (º ´ ½ $ U $ U RÌ » S Ì Ì $U  G³ Î ° (º ´ ½ T U T U RÌ » F Ì Ì TU  G³ Î ° (º ´ ½ S U S U YÌ » Q Ì Ì SU  G³ Î ° (º ´ ½ F U F U XÌ » E Ì FU  G³ Î ° (º ´ ½ Q U Q U RÌ » H Ì Ì QU  G³ Î ° (º ´ ½ E U E U •Ì » Ì Ì Ì EU  G³ Î ° (º ´ ½ H U H U Ò Ì » Ì Ì HU  G³ Î ° (º ´ ½ Ì U Ì U V » Ì Ì ÌU Ë G³ Î ° (º ´ ½ PÄ PÄ $ Ò  »$$ Ì $ PÄ Ë G³ Î ° (º ´ ½ PÄ PÄ T Ò  »TT Ì T PÄ Ë G³ Î ° (º ´ ½ IÄ IÄ S Ò  »SS Ì S IÄ Ë G³ Î ° (º ´ ½ IÄ IÄ F Ò  »FF Ì F IÄ Ë G³ Î ° (º ´ ½ RÄ RÄ Q Ò  »QQ Ì Q RÄ Ë Ò  G³ Î ° (º ´ ½ PÄ PÄ E »EE Ì E PÄ Ë G³ Î ° (º ´ ½ IÄ IÄ H Ò  »HH Ì H IÄ Ë G³ Î ° (º ´ ½ Ä Ä Ì Ò  »ÌÌ Ì Ì Ä E­C ³fgÁ³™ ³fgÁ ¬³«fÃÁ³¶e™A ¶™e Ý!   ¶™e 9ˆD€(Ae À t€(ˆ•€(•B77|@98³ Î 7%Ý ´ ³ ² ¬ Â Æ ² ³ ¶® ¶ Æe « ¯ ° A7g ° Lj¤e ´ F`A~|~Ä ª ! ¼¹»º "Wt(•¹ ! ¼¹»º "Wt(•¹ ! ¼¹»º "Wt(•¹ ! ¼¹»º "Wt(•¹ ! ¼¹»º "Wt(•¹ ! ¼¹»º "Wt(•¹ ! ¼¹»º "Wt(•¹ ! ¼¹»º "Wt(•¹ ³º ¶™ (•­ ° ˆe ³º ¶™ (•­ ° ˆe ³º ¶™ (•­ ° ˆe ³º ¶™ (•­ ° ˆe ³º ¶™ (•­ ° ˆe ³º ¶™ (•­ ° ˆe ³º ¶™ (•­ ° ˆe ³º ¶™ (•­ ° ˆe F­ "ˆC x ¥ — 2 2 ´ ³ ² ¬ ‰ z ¥! 0 E € ¥ … !C €C x   ¥ 0  §© Æ ² ³ (w({Az ° g ° ¨C(¤–0F¤„Fw'dFˆ§ dw¤„AXˆ¤e ´ ¶ C¨(¤–0F¤7dF¦¤t''C {¤#XÇ7ƒƒƒn¨w‰¥ 5‡C D¤„¤t¦F¤R§ ¨¤rƒ(n¦ÇwnC ¤rwBFF}§ ¦( § ‰ z ¥! 0 E C €  0 € … C   0 ©  § 6 i ‘ m C x ¥ !  C B 0! 0 — ! C € §  0 ‰ 0 C  ‘ C  C … ¥  0 ¥ C  ! C ! w i‘ ¥ ¥x 2 0 ! C B 0! 0  C ¥ B C — 2 ¥ …   0   § © §  h CC …C z C©  !C ¥   ¥  ! — ‚C C C x C € £†ƒt¨w‰4d31C D¤„¤t— r¨©y‡%4¨w‰'„¤¢(ˆh ¥†£%dC a“’d¨§F C (ƒ‡ˆ‹F wdC FC C •w}C ¤S¦4F†n¨(F„£di œ ƒ)rF (¤¢w¤}F}C ¤rwBF r§ g¢g~¨§w‰¤d“‚FF¤r¥7wxFrv ¥   !  ¥ ! 0 C … h! § … ¥ € w ž—— š ž ' m § C ‰ 0 E C €  0 ¥ C     … ‰ 0 ¥! 0 B  ¥C ¯ ¯ ¯ $ ̮ º ³ ¬Â ¯ $ ® º ³ ¬® ¸ ¶ ™ e Ý ! ¸Â ¸   ¶ ™ e ݸ ® ¶ Æ e « •A&a#b(£”t%Â Ò `7FaÄ µ %ˆ|##" ”Ì µ %³ Î 7%ƒ€@A€t€Ä µ •€e ´ ¶ ª Ʋ ³ … oDd©ƒ— §(w'~¨§w‰orF‡ƒ(gف#¨C(7C 7aF(('¨(F ¨ ž  š ˜ ¨œ ¦ x ¥ … ‰ §m C € C  ‘ C  Ø  ‰ "  § —!  — ¥ € ! § Ô i C ¥ B C — 2 ¥ …  C ¥ ‰ ! C ¥ B   § h© §  h m §   ! — ‚C C C x C €  ‰  ‰ 0©  C ‰ ‰  § … C Ç7¨©y‹%{¨w‰'D§ ‹ˆwEF¤–0C ‹¨©yg(ˆ„A@¤„¨§¥  C (ƒ}ˆrF‡C ¤–0(¨©¤#y§ ‹ˆwhw¥4¨(¦‹’¥ pH Y c e © P H s £ ¡ F%t'vugT |bW ¤¢ D00 „¤ 2„F¤ˆ§ ¨©yB § ! „w¤„(}ˆ#t¥ˆn7¨§F¤„7¨§‹Fy§ „C ¤–0C }C ¤„F #¢Ç~ w(FC u¦†¦`D§ ¨¤0 0!! 0 ! ¥ B m C ¥ 0 z § C E0 €  § 2  ¥ 0! x ‰ C €    ‰ ! ! 0 2C € …§ • i ‰  ! $ 9x j ‰ … § € 2C € !C — 0 —  0 € w i‘m  i 5 6 6 5 … ‰ 0 $ ‰ ¥! 0 C € ¥ ‰  ! C € € ¥ … ‰  ! !  § 2 ! 0 — B § ~F4Fg%¤td¤tF„¦ƒƒa— 777d#¨§w‰¤kyX¨C¨©FF¤RF„wg w(FC ‡F„Fw'g w(FC d(ˆ‡C ¤ty¦$ W HBYW p |s ŒC`‡1A ¯ @ º À µ ³ f g ½ » ¬Â ° ¾ ° ³ f g Æ ³ ™ e ¸Â º ® ³ ™ « ° e Ë 7¸ ° §¸ µ 7³ ´  •´ ¤Æ ¦Ë (£a¸ •Ay½ €(AYÇ» ° ÚAg y½ ³ ° |¬ ª ¯ ˆ¸ © ¸ µ A« ¦° » ´ aaˆA ¸ A€´ 7¤7FŠGqGS µ (7A£`H µ G•7Ä Â °  ڕ((•» µ 7» ±ª ­ ½ ¬  ¸ ¬ ¶ µ ° à £ ¶ g Q $ E ¬ ³ « f à Æ  ³ » e ® ³ ™ « g · º ° º ° 8 8 7 3 ‰‰ 0   § z 0  §¥  0 B ! §m ¥   0 …  § 2 97¢bw¤¢(a¤’¨F¤rFowr•¤k¥ˆRC … C — 2 " C ‰ 0 Ex —  © !! §© 2 z  ! § C C 7a#w¤7y„C C F4•{C F„dˆx  0© w i ‰ ‘  ‚   0 ³ ™ « g · º m §   —  § C € C E0  0©  § #¥£d7¨CdƒC n˜•7A» ° ¤¢((F(SF}ˆ#„d¤#1Ç6 W l64nP H %(s H f SbgT SP |ÏaY f 5 3c sc W P W T H cH 1 2¡ 0 i C   h© § h C € 0 … ¥ C© ¥  C C h0 B ¥ € j i  ‰ C € ¥  ! 7Cˆxg(§ˆ„A‹Fgk'w‹wF‹C „gˆ¤rg¨(kbƒ‘ F¨w¥‡F4wgF ‘  wnt7§a—dt`m¤§„¨§¥F C !(—ƒC‡CF4w¥'¨¥r¨©¥yrCaA£4¨w‰'…4¤0d(ˆ„A‹ˆ¥¦a‹C ”“’gFˆ7wx(‹kj ‚€ C B — 2 ¥    § h© §  h C C x CC …C z C©  !C ¥  C x ¥ z C € i  § €C B ‘ 0 2!C E§ ¥ ”AFD†~C ¤¢4Fˆ#© ¨¢v  ‰ ¥ ‰  B  C    © !! §© !C … C ‰ 0 Ex — !C l¨C(w—Fw(}˜§ SÉC C FyC †¢w¤7yRC ¢• C ¨¤|oÊ¢`‹ÙFa•„A¢• ‰ 0 ԏ v Ô Ø x! C z €© § ( Ø ¥ ¥ B 0 &  C £ 2 z x ¥   )Ùy(wy¤F'kdA{(wFFC ½ S ÇÒ Ì Ò #•H P AuS ŠÒ err9G9E À » ° Aae ´ ™ U •Ç9A•§H ³ F ÌÌ ½  HVHEF ¬ g ™ e f ­ g ³ Ý · A  ¹ Á ¥ Ä E ½ H H Ò #•H P PuS ŠÒ gGr9X|E Ì Ò rrA %Î ¼ © » ° 7•ae ´ #F A•§RÌ ³ S HÌ ½  TSSTÌ FF $ ¬g™e f " ¹Á¥Ä V Q ½ F Q Ò #•S R RÌ £½ grX|rqT •(³ ¤¤´ •(» ° Aae ´ A•§S ³ T E Ò Ì E$ÌTE ™»¬ °´° g³­ ¬g™e f ¹Á¥Ä V ½ F Q Ò #€Ì P XÌ £½  Y9r9Q » ° •7%#e ´ 7™ Å €t! ´ 9A•§Ä XÌ ³ V Q Ò Ì QÌSE$ ¬g™ef ­ ²e ·A ¹Á¥ T Q ½ T F Ò #•F P PÌ £½ grX%G9Q €´ ³ ³ V H Ò Ì E$ÌQ$ ° ¹Á¥Ä A•§F E ½ $ F Ò #•H Ò gXuV ŠÒ  Y9r9Q #ˆp¤t€FڕA(G(•¹ ³ Ì  QÌ ½ QÌ S E $ ² f¼  ¶e ºe ™Â ¼ ¼ ¼ C » º ¹Á¥ T A•§Ä XÌ ½ Ì ÇÒ ÌuS Ò #€Ì ReRÌ £½  Ò r9GqE G» ° Aae ´ ™ U •Ç9A•§H ³ H Ì  E Ò Ì T E Q T C ¬ g ™ e f ­ g ³ Ý · A  ¹ Á ¥ Ä S ½ÇÒ Ì T Ò #AT ReRÌ $ ½ eqGH Ò 9XÌ •¹ Î ˆ ˆ¤#"7"q¤•7BA•§±Ì ³ HÌ E Ì HE FE Á ·× ·¨!×!C·¹C ¹Á¥Ä ÒÌ £½ •Ì T Ò #•F X PÌ £½ PgGqrq9Q » ° Aae ´ ™ Å €|Ç9A•§RÌ ³ EÌ H Ò E  E F E V E ¹ ¬ g ™ e f ­ ² e  · A  ¹ Á ¥Ä F E ½ FRÌ Ì•#AE I PuH ½ P Ò r9rT Ò Ì #ˆp¤t€FÚt7•¹ ̳ QH HÌ E HFT ² f¼  ¶e ºe ™Â ¼ ¹» º ¹Á¥ A•§Ä T ³ à ™ ½ © 7•g Ç° ¹ ¨ ¥¼³Æ»™C ³fg §(AAD¤Á ¥ ¦¼ ÆfÙ» ˆAAC ¯¸¬¶ µ° 㶠ˆa•• ¸ A¤´ 7¤7g Â Ò Ì µ 7ˆ•Í@H µ AÄ Â ¬ ³ « f à Æ  ³ » ¡ $E TQ 9V Ì$ HÌ$ R|GH $ RÌ E ÌQV |G9Q TT TEF rrrH F RÌ Q SVQ r"YÌ TF VET qqXÌ $ AÌ Ì VEÌ qR|Q V XÌ Q SFE r9GF Ò F TH Q $S VQÌ 9XH Ë » Î U A¡ Ä»™ ° t&ÚA7•7» ±ª e ® ³ ™ «g · º ° …‘ ³»Ä ƒ•H µ A͏ ¨w‰4d30 C D¤„¤t§ Ó(w 0 C F„IˆuC „C (ƒC w2w‰¤–0F•C ”“’±#p § ƒF‚C Éف#Ž‰ ¥ ¥x 2 ! C B 0! 0 —  x ¥ ! §©©  ! § C C x   ! — ‚ ‰ ¥  !C ¥  6 —  C € C  Ø C ¯ ° &® ´ (¢³ µ °  ª e ³­g¡ e …C x0h ¥!  C©  ¥!0E  ¥C 20 7ˆ¤(wF(€ ‡¤–0F¤¥¤–0Fˆ'£ w2((¤¢Ù#¨C|7¨C(¤–0w¥C ‡C Dr–0(}Fg C yχF¥wd(d§ FI ¤¤„FŠ¤1¨§F¤rwBFF}F}ˆ¤rD§  ‰ — — 0 Ø  ‰ " ‰ z ‰ ! ! § B ‰  B! §m  x C € ¥  C  C C  ! !! C ! 0   0 m §  ¥ 0 ¥ C C € C h 0 B ! C ! § w i C  C x €© 0 C ! §m  x 0 x ¥ ‰ — — 0 2 z  C  C x   ! — ‚ C ‰ ¥  ! C ¥    ‘  x ¥ § x ! 0 C ( 1}7ˆI„#‹onFC yÏX„(wAw2((¤1•¢ˆIC  C (ƒIw2w‰¤–0FAC a“’1r§ ˜(wˆSC ¤y’¥ £ sW PW QWssWY q p h f HcPWYW bRbdT bt|b%S‹@W Ff ge %dba`XV H ¤¡ WORKING WITH LIMMA AND AFFYMETRIX DATA PRELIMINARY OPERATIONS A. Start RGui; B. Change the working directory choosing the folder where the Affymetrix data is located; C. Load the workspace you have saved from the AFFY exercise: Load(“Affy_Preprocessing.Rdata”) C. Load the needed libraries: library(affy) library(hgu133aEG1000) library(limma) In this exercise we will perform the following tasks: 1. building the design matrix and defining the contrasts of interest; 2. fitting the lnear model; 3. exploring and writing the results; 1. BUILDING THE DESIGN MATRIX AND DEFINING THE CONTRASTS OF INTEREST design <- model.matrix(~0 + pd[,1]) colnames(design) <- c("NORMAL", "RCC") contr <- makeContrasts(NORMAL-RCC, levels=design) 2. FITTING THE LINEAR MODEL fit <- lmFit(datrma, design) fit2 <- contrasts.fit(fit, contr) fit2 <- eBayes(fit2) 3. EXPLORING AND WRITING THE RESULTS First, we visualize the number of significant genes by Venn diagrams. Here we use p-value < 0.01 and Benjamini-Hochberg p-value correction. vennDiagram(decideTests(fit2, p.value = 0.01, adjust.method = "BH")) we extract the gene symbol and the entrez gene information from the annotation package gs <- as.data.frame(unlist(as.list(hgu133aEG1000SYMBOL))) eg <- as.data.frame(unlist(as.list(hgu133aEG1000ENTREZID))) annot <- cbind(rownames(gs), gs, eg) colnames(annot) <- c(“ID”, “Gene Symbol”, “Entrez Gene ID”) we store the significant genes with their annotation in a data-frame object results <- topTable(fit2, coef=1, n = 1838, genelist=annot) Finally, we save the table of significant genes into a TAB-delimited text file (.txt) write.table(results, "results.txt", sep="\t") FINAL OPERATIONS A. Save the workspace save.image("Affy_Limma.Rdata") B. Save the history savehistory("Affy_Limma.Rhistory") Dario Greco Institute of Biotechnology - University of Helsinki Building Cultivator II, room 223b P.O.Box 56 Viikinkaari 4 FIN-00014 Finland Office: +358 9 191 58951 Fax: +358 9 191 58952 Mobile: +358 44 023 5780 Email: dario.greco@helsinki.fi FINDING OVER-REPRESENTED GO FAMILIES PRELIMINARY OPERATIONS A. Start RGui; B. Change the working directory choosing the folder where the Affymetrix data is located; C. Load the workspace you have saved from the AFFY-LIMMA exercise: Load(“Affy_Limma.Rdata”) C. Load the needed libraries: library(GOstats) library(limma) library(affy) library(hgu133aEG1000) In this exercise we will perform the following tasks: 1. creating the parameters for the hypergeometric test; 2. performing the hypergeometric test; 3. exporting the results. 1. CREATING THE PARAMETERS FOR THE HYPERGEOMETRIC TEST First, we need to find the column of the limma results data frame containing the Entrez Gene Ids (It should be the column 3): colnames(results) Now, we can create the parameters for running the Fisher's Exact Test: params <- new("GOHyperGParams", geneIds = as.vector(results[,3]), annotation = "hgu133aEG1000", ontology = "BP", pvalueCutoff = 0.05, conditional = FALSE, testDirection = "over") In this command, we specify the Entrez Gene Ids, the annotation package, the ontology that we want to assay (BP, MF, or CC), the p-value cut off (here we chose 0.05), whether we want to run a conditional test, and the test direction, for finding the over- or the under-represented families (here we want to find the over-represented families). 2. PERFORMING THE HYPERGEOMETRIC TEST BPover <- hyperGTest(params) 3. EXPORTING THE RESULTS In order to save the results into a data.frame object: BPresults <- summary(BPover) Now we can export the results into a TAB delimited text file: write.table(BPresults, "BP_over.txt", sep="\t") FINAL OPERATIONS A. Save the workspace save.image("Affy_GOstats.Rdata") B. Save the history savehistory("Affy_GOstats.Rhistory") Dario Greco Institute of Biotechnology - University of Helsinki Building Cultivator II, room 223b P.O.Box 56 Viikinkaari 4 FIN-00014 Finland Office: +358 9 191 58951 Fax: +358 9 191 58952 Mobile: +358 44 023 5780 Email: dario.greco@helsinki.fi Clustering – Exercises This exercise introduces some clustering methods available in R and Bioconductor.This exercise uses the prenormalized yest dataset. 1. Reading the prenormalized data Read in the prenormalized Spellman’s yeast dataset: > d<-read.table("combined.txt", sep="\t", header=T, row.names=1) We want only the cdc15 data, so take only those columns from the data: > names(d) > da<-data.frame(d[26:49]) Remove missing values from the data: > dat<-na.omit(da) 2. Filter the genelist by standard deviation Select only the genes that are among the 0.3% of the highest standard deviations. > > > > > > > library(genefilter) # Row-wise SDs sds<-rowSds(dat) # Which is the value at 99.7% of data sdt<-quantile(sds, 0.997) sel<-(sds>sdt) set<-dat[sel, ] How many genes are left after filtering? 3. Creating a heatmap using Euclidean distance and complete linkage > heatmap(as.matrix(set)) To get other colors in the heatmap, you first need to generate a sequence of colors, and then plot the heatmap using these colors: > library(RColorBrewer) > heatcol<-colorRampPalette(c("Red", "Green"))(32) > heatmap(as.matrix(set), col=heatcol) 4. Saving the heatmap into a file For further modifications, the heatmap might need to be saved in a file. This is accomplished with: > > > > cwd=getwd() bmp(file.path(cwd, "heatmap.bmp"), width=1800, height=1800) heatmap(as.matrix(set), col=heatcol) dev.off() This results into about 6*6 inch print quality bitmap image in your data folder. Some papers might want to get a postscript image, and this is accomplished as: > cwd=getwd() > postscript(file.path(cwd, "heatmap.ps"), width=1800, height=1800) > heatmap(as.matrix(set), col=heatcol) > dev.off() 5. K-means clustering of genes In K-means clustering you need to pick an artificial number, the number of clusters (K). To produce a K-means clustering with 5 clusters, type: > k<-c(5) > km<-kmeans(set, k, iter.max=1000) Calculate an average withinness of the results. This is a measure of how close together genes lie inside the clusters. > mean(km$withinss) [1] 21.1838 Run the same K-means analysis several times (save the result into a new object every time). Select the K-means clustering giving the smallest withinness score as the best result. You can do this by hand, or run the following code chunk: > ss<-c(1000000) > for(i in 1:10) { > km<-kmeans(set, 5) > if(mean(km$withinss)<=ss) { > ss<-mean(km$withinss) > km.best<-km > } >} 6. Visualizing the K-means clustering Let’s produce a new K-means clustering result using four clusters: > km<-kmeans(set, 4, iter.max=1000) Next, initiate a 2*2 image area, and draw the expression profiles. We need to apply a for-loop here: > par(mfrow=c(2,2)) > for(i in 1:4) { > matplot(t(set[km$cluster==i,]), type="l", main=paste(“cluster:”, i), ylab=”log expression”, xlab=”time”) >} ...
View Full Document

This note was uploaded on 06/06/2011 for the course STAT 4360 taught by Professor Park during the Spring '11 term at University of Georgia Athens.

Ask a homework question - tutors are online