Deep space, data analysis and dealing with uncertainty

- 1 Comment

In July Source Code for Biology and Medicine, a forward-looking bioinformatics journal with an emphasis on the practical utilization of source code within biomedicine, attended the Ninth International Meeting on Computational Intelligence Methods for Bioinformatics and Biostatistics (CIBB) in Houston, Texas. This meeting took place at the Methodist Hospital Research Institute (TMHRI), which is one of the many hospitals in the renowned Texas Medical Center.

Texas Medical Center is internationally recognized for such prestigious institutions as Baylor College of Medicine, the University of Texas MD Anderson Cancer Center and the Methodist Hospital System. These institutions provide some of the world’s most advanced treatment available and also carry out ground-breaking research. The conference covered a wide range of areas, including medical informatics, genetics, structural biology and radiation biology. A central theme throughout was biostatistics and data analysis methods, particularly those for managing uncertainty within statistics.

The first plenary speech was given by the enthusiastic and eminent IEEE life fellow, Jim Bezdek, who although retired, has seemingly boundless amounts of energy. He is also one of the first and most prominent researchers of fuzzy clustering – a growing area in data analysis. Fuzzy clustering is a group of algorithms used for cluster analysis; the allocation of data points to different groups or ‘clusters’ is not “hard” (all-or-nothing) but “fuzzy”. Fuzzy clustering also allows researchers to more effectively manage uncertainty and noise within data. Jim’s presentation gave a very informative and brief history of the cluster heat map, which is a very popular visual data analysis method in the bioinformatics community. Interestingly, his presentation was also peppered with anecdotes about his fishing exploits (see image right), which I’m sure had some sort of connection to the data analysis methods he was explicating.

Some extraordinary research came from a group of NASA researchers who are one of the few groups in the world to be researching the effect of space radiation on humans and the possible risks associated with this. Other research featured a range of novel methods and algorithms; many are novel methods and algorithms based around fuzzy clustering which ultimately allow a more effective analysis of what are extremely complex and large amounts of data. Some, for example, can be used to predict membrane protein types to facilitate drug discovery, and others can simulate radiation tracks that can assist in radiotherapy treatment planning.

Example of a cluster heat map (Ovaska et al. BioData Mining 2008 1:11)

It was clear to see that progress was made in the fields of biostatistics and bioinformatics at this developing and emergent conference. Furthermore, although much of the research would be somewhat lost on the layman, it was also easy to get a sense of how important such research is to the growth of knowledge and enlightenment in biomedicine; more importantly it was clear how valuable it is to the advancement of health care. It was a pleasure to meet this welcoming and close-knit group of scientists who are so passionate about the research in which they’re involved. In reflection of the broad scope of CIBB, Source Code for Biology and Medicine is looking for manuscripts which exemplify valuable research across the fields of Bioinformatics and Biostatistics. Submit your manuscript here, or contact the Editors-in-Chief.