Course Content
Although many of the topics are generic, this course will be particularly useful for researchers in molecular biology, biochemistry and biotechnology and professionals in pharma-related industries.
Introduction
Why study biostatistics? Relationship between a population and a sample
Descriptive Statistics
Measures of central tendencies (mean, median, mode, percentiles), variation (range, standard deviation, variance, IQR, CV, MAD) and shape (skew, kurtosis) of distributions. Elements of data visualization and exploration.
Concept of probability distributions. Sampling distribution of means. Standard errors. Confidence intervals.
Inferential Statistics
Introduction to hypothesis testing, p-value and its interpretation, One sample t-test, Type I and Type 2 errors.
Power analysis, sample size calculation.
Comparing two means: t-tests (paired, unpaired, Welch's correction).
Common mistakes in statistics I: Lack of proper randomization, pseudoreplication, misinterpretation of p-values and the replication crisis in science.
Going beyond simple p-values: Effect sizes and confidence intervals.
Comparing multiple means I: Concept of Family-wise error rates, and controlling for them through various corrections (Bonferroni, Holm-Sidak stepdown).
Comparing multiple means II: One-Way Analysis of Variance (ANOVA), Comparing various post-hoc tests and recommendations about when to use which test.
Outlier detection: Dixon's Q tests, Grubb's tests, Rosner; Extreme Studentized Deviate Test, Hampel's Test.
Checking the assumptions of parametric tests: Various formal and informal tests of Normality and Homoscedasticity, and commenting on their reliability/applicability.
When assumptions are not met (Part 1): Transformations (when to use which one and when not to).
When assumptions are not met (Part 2): Non-parametric tests for central tendencies (Mann-Whitney U-test, Wilcoxon Rank sum test, Kruskal Wallis test).
Relationship between variables
Contingency Table Analysis, Fisher's exact Test, Chi square test.
Chi square test of goodness of fit.
Correlation analysis (Pearson's, Spearman's).
Linear Regression, Residual analysis.
Multivariate Analysis
Principal Component Analysis (PCA): Concept, Data Interpretation, Scree Plots, Biplots, Dos and Don'ts, Testing of Assumptions
Final Thoughts
Common mistakes in statistics II: A list of various common mistakes and how to avoid them.
Keys for picking up the correct statistical tests.
A 10-point schema for successful data analysis.
Points to note:
Throughout the course, we shall talk about how to report the results of various tests in the literature.
We shall also refer to various recommended best-practices in various fields, and appropriate references shall be provided for the same.
For each topic discussed in the class, the students will be provided with appropriate reading material, to further their knowledge/understanding.