Download 4 Questions on Normalization in Bioinformatics | BCB 444 and more Lab Reports Bioinformatics in PDF only on Docsity! BCB 444/544 Fall 07 Nov. 15 Lab 11 p. 1 Name _____________________________ Email: 1) In your own words, describe in several sentences what normalization is, and why we need to do it prior to any further processing of the data. 2.1) How many gene expression patterns are in the data set? 2.2) How many of them are replicated? 2.3) Are there a lot of missing values, in total? 2.4) Was the data set already in a symmetrical scale? 2.5) How many gene expression patterns have 3 missing values? 2.6) How many genes will remain if we want to keep only genes with at least 3 peaks higher than 1 or lower than -1 (in a log2 scale)? 2.7) Which functions does the pre-analyzer recommend activating? 3.1 In Scenario I, a single number can be changed for Gene 1 that results in dramatic changes in the correlation. Use the two graphs for the scenario to guide your experimentation of the following changes. a. Change a single sample for Gene 1 that causes the correlation to jump up to approximately 0.68, and list the change made. b. Change Scenario I, Gene 1, Sample 8, from 100 to –150. Note that the correlation jumps down to approximately –0.63. Explain this jump by seeing what changes in each of the two graphs. 3.2 To help answer the following, first notice that in Scenario II, the pattern for Gene 2 is evenly spaced between 10 and 80, changing in increments of 10. a. Change the pattern for Gene 1 in Scenario II such that the correlation is exactly 1, and list the changes made. You will need to change all but one or two of the values. b. Change the pattern for Gene 1 in Scenario II such that the correlation is exactly –1, and list the changes made. You will need to change all but one or two of the values. 3.3 Scenario III illustrates how sensitive the correlation can be to small changes. Here we examine a gene whose log ratio changes substantially across samples and a gene with essentially constant log ratio across samples. a. Find a pair of samples for which Gene 2 can be changed from 7 to 6, resulting in a much larger positive correlation, and list the changes made. b. Return the two samples found in part (a) to their original values of 7, and find a new
