"Significant (P < 10e-12) improvements to this work resulted from the reviewers' comments."

Perez-Enciso M. In silico study of transcriptome genetic variation in outbred populations.
Genetics. 2004 Jan;166(1):547-54.

Loengukursused -> Matemaatiline statistika ja modelleerimine


Mathematical statistics and modelling
[Matemaatiline statistika ja modelleerimine]

(DK.0007; 5 EAP; spring; E)
EMÜ Doctoral School


Course info Lectures Practicals

Exam

Links

NB! The deadline for individual works: April 28, 2013

 

Exam consists 3 parts: individual home works (40% of points), project/presentation (30%) and test (30%).

1. Individual home works

DK_exercises_2013_eng.pdf


2. Project or presentation about own research:

  • research area; interested questions, problems, hypothesis;

  • how is planned to perform the experiments (or how the experiments were/are perfomed), collect data, make questionaries, measure something, ...;

  • how many experiments/studied objects, why so many/few, how selected, ...

  • how should these data be analysed (or discussion about this, what kind of results are requested to answer the questions, solve the problems, ....
  • ...

3. Test

To perform the test successfully, you should understand

  • how are calculated basic descriptive characteristics like aritmetic mean (average), median, standard deviation, variance, quartiles; when is preferred average and when median; how differ mean and median in case of asymmetric distribution;
  • what descibes/measures standard error, confidence interval, significance level, p-value;
  • how looks like the density function of normal distribution, what are the basic properties of the normal distribution;
  • when it is suitable to use the t-test, Mann-Whitney or Wilcoxon test, sign test, F-test, Kolmogorov-Smirnov test, chi-square test, Fisher exact test; what is the Bonferroni correction/method;
  • for what kind of analyses should be used two-way frequency tables, Pearson and Spearman correlation coefficients, linear regression, analysis of variance (ANOVA), general linear models (GLM);
  • when are used logistic models (logistic regression), probit-models (probit-regression); what is the odds ratio (OR), how is the odds ratio related with the two-way frequency table or logistic regression;
  • what is measuring the model standard error, multiple correlation coefficient, determination coefficient (R2);
  • what is the meaning of linear model's parameters (regression coefficient, parameters of analysis of variance model), what are the contrasts;
  • for what kind of analyses should be used multivariate analysis methods like cluster analysis, principal component analysis, discriminant analysis.