In silico microRNA profiling for detection of cancers /
Biopsy examinations for the purpose of cancer detection and classification are highly invasive since tissue samples are required to be extracted from patients. Upon diagnosing cancer, specific clinical treatments have to be sought based on the class of cancer detected. Therefore, cancer classificati...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
Kuala Lumpur :
Kulliyyah of Engineering, International Islamic University Malaysia,
2018
|
| Subjects: | |
| Online Access: | Click here to view 1st 24 pages of the thesis. Members can view fulltext at the specified PCs in the library. |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Biopsy examinations for the purpose of cancer detection and classification are highly invasive since tissue samples are required to be extracted from patients. Upon diagnosing cancer, specific clinical treatments have to be sought based on the class of cancer detected. Therefore, cancer classification is imperative. Currently, a few invasive cancer biomarkers in forms of proteins from serum are available. Nevertheless, existing cancer diagnosis techniques require labor-intensive analysis compounded by low diagnosis sensitivity. There have been a number of studies to identify non-invasive novel miRNA-based cancer biomarkers. However, the existing non-invasive techniques using miRNA suffer from low diagnosis accuracy, sensitivity, and specificity. This thesis puts forward a machine learning based framework that could recognize and classify cancer from miRNA expression data. To circumvent the problems faced by the state-of-the-art cancer classification systems, this work employs an entropy-based transcriptomic marker selection approach to select relevant miRNA markers that are directly responsible for cancer discrimination. To investigate the involvement of miRNAs in cancers, experimentally validated and computational predicted target genes of miRNA biomarkers were collected from repositories such as miRWalk and TargetScan. Eight challenging cancer datasets, namely meningioma, breast cancer, diffuse large β-cell lymphoma, schwannoma, prostate cancer, ovarian cancer, chronic lymphocytic leukaemia and multiple myeloma from the ArrayExpress database were selected for this study. However, miRNA network related to breast, prostate and ovarian cancers were analyzed because these three cancers are the most common hormone-related cancer types and facing high metastatic rate as well as poor prognosis outcomes. Besides, the miRNA regulatory mechanisms associated with these three cancer types are still unclear. The proposed cancer classification system managed to acquire 100% Leave One Out Cross Validation (LOOCV) accuracy rates for four out of eight datasets. The mean LOOCV accuracy rates across all the eight dataset is 98.60%. The proposed system was found to outperform the state-of-the-art systems by circumventing the fundamental problems caused by the state-of-the-art systems. Additionally, proposed system able to identify novel miRNA biomarkers for five out of eight cancer datasets. The results demonstrate the efficacy of the proposed cancer classification framework. Gene ontology analysis results exhibit that the identified miRNA biomarkers of breast, prostate, and ovarian cancers were found to be implicated in regulation of major cancer-related biological processes. Functional analysis results show that these miRNA biomarkers were found to target numerous cancer initiating genes. Precisely, miR-125b, was found to target the largest quantity of genes coding oncogenic transcription factors. Specifically, miR-145 was found to target the highest number of genes coding nucleic acid binding proteins. Pathway analysis results also exhibit that the identified miRNA biomarkers of breast, prostate, and ovarian cancers were found to interact with major cancer associated signaling pathways such as MAPK, WNT, TGF-beta, mTOR as well as secondary cancer associated signaling pathways such as androgen receptor signaling pathway and Toll-like receptor signaling pathway. The proposed framework can be applied in various areas of clinical oncology such as cancer diagnosis, cancer clinical outcome classification, and cancer clinical grade and stage classification. |
|---|---|
| Physical Description: | xxi, 211 leaves : illustrations ; 30cm. |
| Bibliography: | Includes bibliographical references (leaves 142-168). |