Publications

Publications

Reporting and analyzing alternative clustering solutions by employing multi-objective genetic algorithm and conducting experiments on cancer data

By:
Contributors: Reda Alhajj, PhD, Mohamad Elzohbi, Peter Peng

Knowledge-Based Systems 56 (2004) 108-122

Peter Penga, Omer Addama, Mohamad Elzohbia, Sibel T. Özyerb, Ahmad Elhajjc, Shang Gaoa, Yimin Liua, Tansel Özyerd, Mehmet Kayae, Mick Ridleyc, Jon Roknea, Reda Alhajja, f

Abstract

Clustering is an essential research problem which has received considerable attention in the research community for decades. It is a challenge because there is no unique solution that fits all problems and satisfies all applications. We target to get the most appropriate clustering solution for a given application domain. In other words, clustering algorithms in general need prior specification of the number of clusters, and this is hard even for domain experts to estimate especially in a dynamic environment where the data changes and/or become available incrementally. In this paper, we described and analyze the effectiveness of a robust clustering algorithm which integrates multi-objective genetic algorithm into a framework capable of producing alternative clustering solutions; it is called Multi-objective K-Means Genetic Algorithm (MOKGA). We investigate its application for clustering a variety of datasets, including microarray gene expression data. The reported results are promising. Though we concentrate on gene expression and mostly cancer data, the proposed approach is general enough and works equally to cluster other datasets as demonstrated by the two datasets Iris and Ruspini. After running MOKGA, a pareto-optimal front is obtained, and gives the optimal number of clusters as a solution set. The achieved clustering results are then analyzed and validated under several cluster validity techniques proposed in the literature. As a result, the optimal clusters are ranked for each validity index. We apply majority voting to decide on the most appropriate set of validity indexes applicable to every tested dataset. The proposed clustering approach is tested by conducting experiments using seven well cited benchmark data sets. The obtained results are compared with those reported in the literature to demonstrate the applicability and effectiveness of the proposed approach.

Download PDF

 

Stay Informed

To stay up to date on all the latest news and publications, subscribe to our newsletter!

Why did the chicken cross the road…?

Konstantin Stoletov and Lian Willetts co-first-authored an article published recently in Nature Communications titled “Quantitative in vivo whole genome motility screen reveals novel therapeutic targets to block cancer metastasis“. These two researchers, along with fellow Lewis lab members and collaborators from the University of Calgary and Vanderbilt University set out to determine what genes and signaling networks are involved in the rate-limiting steps of solid tumour cell motility, in vivo. But the team was hampered by the lack of an effective, quantitative, in vivo imaging platform. They wanted to visualize the movement of tumour cells, or lack of, in real-time AND use this intravital imaging platform to screen a large bank of tumour cells harboring single gene mutations for cells that show a loss of motility.

https://www.nature.com/articles/s41467-018-04743-2

When tumour cells metastasize they get into (intravasate) the hosts’ bloodstream and use the vascular system like roadways to travel throughout the body. This lets the tumour cells colonize new microenvironments where they will proliferate and form new tumours. So metastatis is really dependent on tumour cell motility. Although there are many different types of solid tumours known, previous research suggests that if the tumour cells can mobilize and metastasize then the expressed motility-related genes share homology across tumour types. This is great news because it would mean that therapeutic targets aimed at stopping motility could also stop metastasis for many tumour types!

Dr. Konstantin Stoletov

 

Dr. Lian Willetts
Dr. Lian Willetts

The Lewis lab researchers and their collaborators developed an in vivo, fluorescent, time-lapse screening platform that uses shell-less avian embryos for tumour growth and formation. The avian embryo is an excellent tumour model because the tumour cells will grow on the chorioallantoic membrane in a single cell layer, making in vivo cell motility imaging actually doable.

Using this platform the team screened over 30 000 human genes for the ones needed for cell motility and ultimately found 17 genes that looked to be effective metastasis-blocking gene targets. Stoletov, along with other Lewis lab members, are continuing this research by studying these 17 attractive candidates further to determine which one (s) would make therapeutic metastasis-blocking targets.

This article has generated a lot of interest in the scientific community and in the general public! Check out the links below to mentions and articles in the media.

https://www.biocentury.com/bc-innovations/translation-brief/2018-07-18/how-chicken-embryo-screen-identified-entos%E2%80%99-

https://www.ualberta.ca/medicine/news/2018/june/putting-the-brakes-on-metastatic-cancer

Stay tuned for a podcast that will be posted soon from “Parsing Science” where the hosts interview Dr. John Lewis about this work!

https://www.parsingscience.org/coming_soon/
UPDATE Oct 12, 2018: The podcast with John Lewis on Parsing Science called “Halting Cancers’ Spread“, is now available!

- Perrin Beatty