Prediction of Beta Lactamase using Position and Composition Variant Features and Neural Networks

Simple item page
dc.contributor.authorAdeel Ashraf, Muhammad
dc.date.accessioned2018-02-23T12:31:46Z
dc.date.available2018-02-23T12:31:46Z
dc.date.issued2017
dc.descriptionSupervised by: Dr. Yaser Daanial Khanen_US
dc.description.abstractβ-lactamase produced by different bacteria confers resistance against β-lactam containing drugs. The gene encoding β-lactamase is plasmid borne and can easily be transferred from one bacterium to another during conjugation. By such transformations, the recipient also acquires resistance against the drugs of β-lactam family. β-lactam antibiotics play an important role in clinical treatment of diseases like skin infections, soft tissue infections, urinary tract infections and bronchitis. Furthermore, based on the physical structure of beta-lactamases, they can be classified into four classes namely A, B, C, and D. Class A, C and D include enzymes that hydrolyze their substrate by forming an acyl enzyme through active site serine while class B are metallo-enzymes that utilize at least one active site zinc ion to hydrolyze. This Thesis presents a computationally intelligent technique formulated to predict whether a given protein belongs to beta-lactamase or non-beta-lactamase. If the given sequence is beta-lactamase then identify beta-lactamase to its respective class (Class A, B, C, D). The computational model uses primary amino acid sequence structure as its input. Various metrics are derived from the primary structure to form a feature vector. Experimentally determined data of positive and negative beta-lactamases are collected and transformed into a feature vector space. An adaptive training algorithm is used to train a back propagation neural network for prediction purposes. Rigorous verification and validation tests are performed and metrics are collected to evaluate the authenticity of the proposed model. To differentiate beta-lactamase from non-beta-lactamase with an accuracy of 89.1% and Mathew correlation coefficient of 0.7823. Tests also reveal that the proposed model yields of 94.1%, 95.2%, 93.8%, 97.7%, and 90.4% along with Mathew correlation coefficient of 0.64, 0.75, 0.84, 0.73, and 0.81 for classes A, B, C, D and non-beta-lactamase class respectively.en_US
dc.identifier.urihttps://escholar.umt.edu.pk/handle/123456789/2738
dc.language.isoenen_US
dc.publisherUniversity of Management and Technologyen_US
dc.subjectBetal lactamaseen_US
dc.subjectNon beta lactamaseen_US
dc.subjectMS Thesisen_US
dc.titlePrediction of Beta Lactamase using Position and Composition Variant Features and Neural Networksen_US
dc.typeThesisen_US
Files
Original bundle
Now showing 1 - 2 of 2
Thumbnail Image
Name:
Summary.pdf
Size:
620.77 KB
Format:
Adobe Portable Document Format
Description:
Download
No Thumbnail Available
Name:
Full View.htm
Size:
23.33 KB
Format:
Hypertext Markup Language
Description:
Download
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.71 KB
Format:
Item-specific license agreed upon to submission
Description:
Download
Collections
Department of Computer Science