Term
| Which test imply distribution? |
|
Definition
Parametric tests (t-test, ANOVA)assume a distribution (usually that errors are normally distribted)
Non-parametric tests (Mann-Whitney U, Wilcoxon rank test) analyse ranks, i.e.assume no distribution |
|
|
Term
| What T-test compares? What is the null hypothesis? |
|
Definition
• Compares means of samples
• Null: means are the same |
|
|
Term
| What is the difference between two-sided and one-sided T-tests? |
|
Definition
| Two-sided: means are different, one-sided - higher or lower |
|
|
Term
| What one-sample T test tests? |
|
Definition
The one-sample t-test is used only for tests of the sample mean.
uncertainty is only in one sample |
|
|
Term
| What two-sample T test tests? |
|
Definition
| Two-sample: uncertainty is in two samples |
|
|
Term
| What are two variants of two-sample T test tests? |
|
Definition
– Variant1: both samples have the same variance
– Variant 2: two samples have different variances |
|
|
Term
| What paired-sample T test tests? |
|
Definition
You use the paired t-test when there is one measurement variable and two nominal variables.
The most common design is that one nominal variable represents different individuals, while the other is "before" and "after" some treatment. |
|
|
Term
|
Definition
| Standard deviation. The SD quantifies scatter — how much the values vary from one another. |
|
|
Term
|
Definition
| Standard error of mean.The SEM quantifies how precisely you know the true mean of the population. It takes into account both the value of the SD and the sample size. |
|
|
Term
| 95% of confidence interval (CI) of the mean = ? |
|
Definition
| mean +- 1.96 SEM for a normal distribution |
|
|
Term
| Which types of plots should we use for non-paramentic tests and why? |
|
Definition
| Box-plots, because they imply no distribution |
|
|
Term
| What is familywise error rate? |
|
Definition
| FWER is the probability of making one or more false discoveries, or type I errors, among all the hypotheses when performing multiple hypotheses tests. |
|
|
Term
|
Definition
| if it is desired that the significance level for the whole family of tests should be (at most) α, then the Bonferroni correction would be to test each of the individual tests at a significance level of α/n |
|
|
Term
| Problem of Bonferoni correction |
|
Definition
• Problem : loss of power, especially if there are many true positives and many tests, but few replicates
• Solution: determine the False Discovery rate that controls the number of false positives among all discoveries |
|
|
Term
| What is false discovery rate? |
|
Definition
| Fraction of false positives among all discoveries |
|
|
Term
|
Definition
| The q-value of an individual hypothesis test is the minimum FDR at which the test may be called significant. |
|
|
Term
| What are the methods of gene clustering? |
|
Definition
Pairwise distances =>
Hierarchical clustering (b)
K-means clustering (c)
Self organizing maps (d) |
|
|
Term
|
Definition
| Put the individual P from smallest to largest. Then compare each individual P-value to (ind/m)Q, where m is the total number of tests and Q is the chosen false discovery rate. The largest P-value that has P<(ind/m)Q is significant, and all P-values smaller than it are also significant. |
|
|
Term
| PINGO: You have measured gene expression levels of a gene in blood cells of 10 humans, each before and after eating. Which type of test would be most appropriate? |
|
Definition
| A paired sample t-test, comparing the difference before and after eating for all 10 humans |
|
|
Term
| PINGO: In a paper it is written that Gene X is 2.3-fold differently expressed between patients (mean=8.8, sd=0.56, N=5) and controls (mean=7.7, sd=0.37, N=5), which is significant (t-test, p=0.005).Which statements are correct? |
|
Definition
Mean expression levels are given in log2 space
The standard error of the mean for patients is 0.25 |
|
|
Term
| PINGO: Which statements regarding the False Discovery Rate (FDR) are true? |
|
Definition
+The FDR is used to correct for multiple testing
+The Benjamini-Hochberg procedure is a common way to calculate the FDR
+Storey q-value is a common way to calculate the FDR |
|
|
