Download Stat 371: Introductory Applied Statistics for the Life Sciences and more Study notes Statistics in PDF only on Docsity!
Stat 371
Introductory applied statistics
for the life sciences
Karl Broman
Department of Biostatistics and Medical Informatics
Offices: 6763 MSC, 1274 Genetics-Biotechnology Email: kbroman@biostat.wisc.edu
http://www.biostat.wisc.edu/~kbroman
TA : Lili Lan (lan@stat.wisc.edu)
Logistics
Lectures: MWF 9:55-10:45 (1240 Comp Sci & Stat)
Discussions: 331: T 1:20p (6101 Soc Sci)
332: T 2:25p (4308 Soc Sci) 333: T 3:30p (1289 Comp Sci & Stat) 334: M 12:05p (138 Psychology)
Office hours: Karl : Mon 1-2, Thu 1-2, Thu 3-4 (6763 MSC) or by appointment
Lili : Fri 11am–1pm (B248-H MSC)
Textbooks: Samuels & Witmer (2002) Statistics for the life sciences [required] Gonick & Smith (1993) The cartoon guide to statistics. [recommended]
Grading
Grade based on:
- Homework assignments (25%)
- Midterm 1 (20%)
- Midterm 2 (20%)
- Final (35%)
Other work:
- Reading assignments
- Play with the R software
- Deep and careful thought
- Discussions
3
Computer package: R
Advantages
Free
Available for Windows, Mac OSX, Unix
Comprehensive
Powerful graphics
Well-designed programming language
Unlimited extensibility
Widely used by statisticians
Increasingly used for microarray analyses
Disadvantages
- No dedicated support
- Complex syntax
- Not point-and-click
- Some simple tasks are rather hard
What is statistics?
- Data exploration and analysis
- Inductive inference with probability
- Quantification of uncertainty
7
Example 1
In a study of nephroblastoma (embryonic kidney cancer) in rats exposed to N-ethyl-N-nitrosourea (ENU):
F344 strain: 0%
Noble strain: 50%
Is this difference real?
Example 1
In a study of nephroblastoma (embryonic kidney cancer) in rats exposed to N-ethyl-N-nitrosourea (ENU):
F344 strain: 0%
Noble strain: 50%
Is this difference real?
What if the data are as follows?
F344 strain: 0/
Noble strain: 2/
9
Example 1
In a study of nephroblastoma (embryonic kidney cancer) in rats exposed to N-ethyl-N-nitrosourea (ENU):
F344 strain: 0%
Noble strain: 50%
Is this difference real?
How about these data?
F344 strain: 0/
Noble strain: 8/
0 5 10 15 20 25 30 35
0
20
40
60
80
Quinine conc (ppb)
Fluorescence intensity
13
Example 3
[Carroll, J Med Entomol 38 :114–117, 2001]
Place tick on clay island surrounded by water, with two capillary tubes: one treated with deer-gland-substance; one untreated.
Does the tick go to the treated or the untreated tube?
Tick sex Leg Deer sex treated untreated male fore female 24 5 female fore female 18 5 male fore male 23 4 female fore male 20 4 male hind female 17 8 female hind female 25 3 male hind male 21 6 female hind male 25 2
Example 3 (cont.)
Questions:
- Is the tick more likely to go to the treated tube?
- Do the sex of the tick or deer, or the leg from which the gland substance was obtained, have an effect on the response of the tick?
15
Example 4
For each of 8 mothers and 8 fathers, we observe (estimates of) the number of crossovers, genome-wide, in a set of independent meiotic products.
Question:
Do the fathers (or mothers) vary in the number of
crossovers they deliver?
Example 4 (cont.)
How do we think about this?
If there were no relationship between family ID and number of crossovers in a meiotic product:
- What sort of data would we expect?
- What would be the chance of obtaining data as extreme as what was observed?
19
Fundamental idea # 1
Separate the underlying population from the sample/data.
Separate features of the population (called parameters) from estimates based on the sample/data.
Fundamental idea # 2
Imagine repeating the whole process again.
What other data might we have obtained?
If, in truth, the world is boring, would these data be a surprise?
21
Goals for the course
- Impart the statistician’s view of the world
- Basics of statistics - Basic experimental design - Sampling distributions - Confidence intervals - Hypothesis testing
- Basic statistical graphics
- Basic knowledge of R
