Current Collaborative Research & Development Projects

• Microbial Genome Annotation Interface ("Genotator") with Dr. Steven Gill from the Department of Oral Biology
• Bioinformatics Portal Development serving the University at Buffalo and the Roswell Park Cancer Institute
• Cancer Biomarker Identification in aCGH Data with Dr. Dan Gaile from the Department of Biostatistics
• Cancer Biomarker Identification in Mass Spec Data with Dr. Latif Kazeem from the Roswell Park Cancer Institute
• Reconstruction of Ancestral Spider Silk Sequences Undergoing Gene Conversion and Concerted Evolution with Dr. Katharina Dittmar from the Department of Biological Sciences
• Development & improvement of a microarray overlay tool with Dr. Ken Lo from the Roswell Park Cancer Institute

Research Interests

My main interest is the development of pattern discovery algorithms utilizing multiple data loci to characterize phenotype. This involves a level of combinitorics to be able to define a phenotype (e.g. cancer) in terms of discrete data. Biologocal data is also often flawed, so the methods developed have to be forgiving and intuitive. Pattern discovery is useful in phylogenetics, biomarker identification, detection, data mining, and finding functional linkages, among other things. Most methodologies today are only scraping the surface to find the low-hanging-fruit: find a single canonical discrete piece of data which can be associated to a phenotype. Typically, these methods cannot take into account samples where this one piece of data is missing: not because it wasn't gathered, but because the sample doesn't have it. It also often involves reliance on negative results, driving up false positive rates. If multiple data points are utilized and only positive data yields meaningful results, you can model complex systems more accurately by accounting for more than just one component of it. This model reflects the biological system itself by identifying related components and functions. A change in any component in a system can have the same result on overall function of that system and the system will respond in different ways to make up for it. Thus, using a single data point is only part of the answer. I've worked on a number of projects where my pattern finding algorithms have been employed in strain identification and finding biomarkers associated with a phenotype. Specifically, I helped USAMRIID develop microarrays for a pathogen identification system developed for deployment in all major airports, I helped identify genes associated with virulence for DSTL, and am currently working to identify cancer biomarkers in aCGH and Mass Spec data.

Additionally, I like to employ my biological experience in designing databases and analysis tools for various kinds of biological data. A couple examples would be a tool I developed called genWave which uses the daubechies wavelet transform to find low frequency shifts in genomic sequence ascribed to horizontal transfers in microorganisms and the various microbial databases I've developed for sexually transmitted diseases, oral pathogens, and toxins and virulence factors. However, my main interest is in complex algorithm development.

Publications

1. P Hraber, R Leach, L Reilly, J Thurmond, K Yusim, C Kuiken, "Los Alamos hepatitis C virus sequence and human immunology databases: an expanding resource for antiviral research", Antiviral Chemistry and Chemotherapy 18(3), 113-124 (2007).
2. Paige E. Pardington, Robert W. Leach, Richard T. Okinaka, Paul Keim, and Robert B. Cary. "Francisella Genotyping Microarrays" Paper Pending.
3. R. Leach and T. Brettin, "Generational Difference Analysis" Technical Report, Paper Pending.

Posters

1. R. Leach (2003) The Sexually Transmitted Diseases Database. Frontiers in Bioinformatics, UB Center for Excellence in Bioinformatics.
2. S. Bhotika, R. Leach, et al (2003) Pathogen Databases @ Los Alamos National Laboratory.

Presentations

1. R. Leach (2005) "Toxin & Virulence Factor Database & genWave". At "TSWG Annual Program Review Meeting", Los Alamos National Laboratory.
2. R. Leach (2004) "SNP Extension-based MicroArray Probe Selection". At "TSWG Annual Program Review Meeting", Los Alamos National Laboratory.
3. R. Leach (2004) "Yersinia pestis Pathogenicity Island Analysis". At LANL Meeting, Los Alamos National Laboratory.
4. R. Leach (2004) "SNP Array Design & Analysis & genWave". At "TSWG Review Meeting", Los Alamos National Laboratory.
5. R. Leach (2003) "The TVFac Database". At "LANL / Mitre External Collaborator Meeting", Los Alamos National Laboratory.
6. R. Leach (2003) "Predicting the Expression of a Gene of Interest in Gene Expression Array Data". At "LANL External Collaborator Meeting", Los Alamos National Laboratory.
7. R. Leach (2003) "LANL Sequence Databases". At LANL Meeting, Los Alamos National Laboratory.
8. R. Leach (2003) "ORALGEN: New Functionality of '03". At LANL Meeting, Los Alamos National Laboratory.
9. R. Leach (2002) "Software Configuration Management". At LANL Meeting, Los Alamos National Laboratory.
10. R. Leach (2001) "Multi-Database Interaction Application". At LANL Meeting, Los Alamos National Laboratory.
11. R. Leach (2001) "Presenting Bugspray 2.0". At LANL Meeting, Los Alamos National Laboratory.
12. R. Leach (2001) "The Genome Annotation Toolkit". At LANL Meeting, Los Alamos National Laboratory.

Links

I've worked on a number of bioinformatics projects in my career. Some are no longer online, but those which are still up and running are listed here.

1. TVFac - The Toxin & Virulence Factor Database
2. Oralgen - The Oral Pathogens Database
3. ATK - The Genome Annotation Toolkit
4. STDGEN - The Sexually Transmitted Diseases Database
5. BASIS - The Biological Aerosol Sentry and Information System
6. HCV - The HCV Sequence Database
7. Flu - The Influenza Sequence Database

© Robert W. Leach, 2008