Antibiotics are efficient drugs against infections caused by bacteria. However, bacteria can quickly become resistant against antibiotics which require new drugs to be constantly developed resulting in a process that is associated with high costs. Identification and understanding of antibiotic resistance and how it develops is important to keep the design of new antibiotics feasible. Antibiotic resistance is often caused by the horizontal transfer of genes in which the bacteria can share genetic information between different organisms. Antibiotic resistance can therefore spread rapidly, causing untreatable infections all over the world.
In our group, we have created a probabilistic model, namely a hidden Markov models (HMM), to identify genes that have been horizontally transferred by looking for the so called attC sites, a 59 bases element that mediates the gene transfer. Now, we are interested in compare our methods to other existing ones. It is likely that existing methods will look for structures that include the attC sites but are bigger than them.
The tasks in this project include
- Literature study the existing methods.
- Implement of a framework for cross-validation of methods for identification of attC sites. The framework should calculate sensitivity and specificity for each evaluated method.
- Evaluate and compare the performance of these methods.
An introductory lecture on the related biology and our model will be held in the beginning. Parts of the literature for part 1 will be provided. However, an extension through a literature search will also be necessary. Implementations will be done in R. A comparison summary should be prepared.
Obs! För GU-studenter räknas projektet som ett projekt i Matematisk Statistik (MSG900/MSG910).
Examinator Maria Roginskaya
Institution Matematiska vetenskaper