Establishing experimental criteria for probe design and developing new computer programs

Microarrays fabricated with the genes encoding key, functional enzymes involved in various biogeochemical cycling processes are referred to as functional gene arrays (FGAs). One of the greatest challenges in using FGAs for detecting functional genes and/or microorganisms in the environment is to design oligonucleotide probes specific to target genes/microorganisms because many sequences targeted in environmental studies are highly homologous. To tackle this problem, we have experimentally established the probe design criteria by considering sequence homology, free energy and sequence stretches. This paper was published in Applied and Environmental Microbiology (AEM) and was listed as one of the top 20 papers most requested in 2005 by AEM. We have further improved the probe design criteria by showing that very little cross-hybridization was observed when a probe has ¡Ü 90% sequence identity, ¡Ý -35 Kcal/mol of free energy, and ¡Ü 20 base stretches with its non-targets. Our recent tests with FGA-II (>24,000 gene probes) showed that the probe specificity can be predicted very well based on these criteria.

Once the criteria are established, the next critical issue is the development of a computer program to implement probe design strategies. Due to the highly homologous nature of the gene sequences of interest for community studies, no commercial software or freeware was available. Thus, over the past two years, we have developed, tested, and applied a new software tool (CommOligo) to design oligonucleotide probes for microarrays. There are several advantages for the newly developed software. First, this program was specifically designed by considering the challenges of selecting specific probes based on many homologous sequences of each functional gene category. Thus it is well suited for microbial community study and microbial detection. To our knowledge, this is the first computational program for designing community-wide gene probes. It can also be used for designing probes for whole genome arrays. Second, this program implements novel global alignment algorithms, and simultaneously uses multiple criteria (e.g. sequence identity, hybridization free energy (¦¤G), and continuous stretch to predict oligonucleotide specificity. In addition, the program has a unique feature to design group-specific probes for a group of highly homologous sequences.

For monitoring microbial populations in the environment, it is ideal to develop hierarchical oligonucleotide probes (50mer) for different phylogenetic groups of genes/microorganisms at different levels of specificity (e.g., strains, species, genera, and families), and for individual sequences/organisms as well. Such hierarchical probes will allow us to simultaneously detect both closely and distantly related genes/populations so that we will not miss some important distantly related populations. However, the currently developed program could not meet such requirements. Thus, we are developing computational tools capable of designing hierarchical probes based on the CommOligo program.

Development of comprehensive functional gene arrays (FGAs). Although many technical issues regarding microarray technology have been solved, one of the critical bottlenecks that remain is to design FGAs containing probes appropriate for studying the microbial communities of interest. With the newly developed computer program, we have recently finished the construction of the second generation of FGAs for environmental studies. The probes designed for this microarray encompass the variation in >10,000 known microbial functional genes involved in nitrogen (e.g. nitrification, denitirification and nitrogen fixation), carbon (e.g. carbon dioxide fixation and cellulose degradation, methane production and oxidation) and sulfur (e.g. dissimilatory sulfite reduction) cycling processes, organic contaminant degradation, and metal resistance and reduction. To our knowledge, this is the most comprehensive arrays in the world available for environmental studies. Now, we are developing more comprehensive FGAs by including more functional gene groups.

We have also developed other novel high-throughput genomic technologies such as community genome arrays (CGAs). We have evaluated the specificity, sensitivity, quantification and potential application for environmental samples under a variety of conditions. Our results indicate that microarray-based genomic technology has potential as a specific, sensitive and quantitative tool in detecting microorganisms in environments and revealing species relationships among different bacteria.

Novel approaches for increasing microarray hybridization sensitivity

One of the main challenges in using microarrays to analyze microbial communities in natural settings is current detection sensitivities are not sufficient for detecting the majority of microbial populations in environmental samples. Thus we have initiated the development of new technologies for increasing detection sensitivity. First we have optimized the labeling and hybridization systems. With such optimization, 50-100 fold sensitivity can be increased. Second we have developed nanotechnologies by using nano-particles for increasing signal detection on microarrays. Our results showed that 10-50 sensitivity can be increased. In addition, we have developed a novel approach and strategy for representatively, quantitatively amplifying whole microbial communities for microarray-based detection. With this technology, as low as two bacterial cells can be detected. Application of this technology to various environmental samples showed that this technology can provide reliable and quantitative detection of microbial populations in environmental samples. This is the first time that microarrays have been used to visualize a complete picture of microbial communities in natural settings with low biomass. The development of such technologies makes it possible to utilize microarrays for analysis of environmental samples. This approach will also be very useful for addressing questions concerning microbial communities associated with human health, plant and animal quarantine (e.g., pathogen detection), plant ecology (e.g., rhizosphere populations), animal productivity and health (e.g., intestinal and rumen populations), forestry, oceanography, fisheries, ecology, biodiversity discovery and management (e.g., pharmaceutical discovery), etc., as microbial communities play important roles in each of these areas and the available natural community biomass is often very restricted.

mRNA-based detection in environmental samples. One big problem in the detection of activities in environmental samples by microarray hybridization is also to obtain sufficient amount of mRNAs for analysis. We have developed and evaluated an mRNA amplification strategy for improving detection sensitivity of microbial activity using a modified T7 RNA polymerase-based approach to amplify prokaryotic mRNAs for microarray analysis. Our results indicated that mRNA can be successfully amplified by this approach with community mRNAs and enough mRNAs can be obtained for microarray hybridization from contaminated groundwater samples. This development of such technology makes it possible to analyze microbial community activities in environmental samples.

Novel surface chemistry for microarray fabrication

For constructing microarrays with short oligonucleotides (<25bp), one end of an olignucleotide probe must be modified for attachment to glass slides, which generally costs $5-20 per modification. Such cost is prohibitive when high density of oligo arrays is considered. Thus we have also developed a novel surface coating chemistry for fabricating unmodified oligonucleotides on glass slides. This novel chemistry has significant advantages over the currently used approach in terms of detection sensitivity, dynamic range, versatility, background and cost. Our experimental results showed that single base difference between probe and target DNA can be easily differentiated with the new surface chemistry. Also this type of chemistry can be used for fabricating DNAs and proteins. A patent is in pending and this technology is licensed to a company for commercialization.

Protein arrays

Protein microarrays are becoming an important tool in proteomics, drug discovery and disease diagnosis. Fabrication of protein microarrays on planar surfaces is a great challenge due to protein denaturation and cross-activities. We have developed three-dimensional nanofilm-based slides for protein array fabrication. Our results showed that very reliable antibody-antigen interactions can be detected with the new surface chemistry. We have also developed a novel hydrogel-based approach for protein array fabrication. Our results showed that the activity and selectivity of the hydrogel-based protein arrays are advantageous over conventional technologies. A patent on this technology is in preparation.

New methods for sample preparation

One of the bottlenecks in using microarrays for environmental studies is the sample preparation. Thus, besides the development of microarrays-based genomic technology, my laboratory has also developed and optimized molecular methods for extracting high-quality nucleic acids from environmental samples. We were the first to be able to recover high-quality intact mRNAs and DNA simultaneously from a variety of soils and sediments, which are extremely challenging. My laboratory was also the first to discern the problems of heteroduplexes and PCR-induced mutations in the 16S rRNA gene-based cloning approach and to develop approaches to minimize such artifacts. These studies are important in molecular microbial ecology because PCR-based cloning approach is the most widely used and powerful one for analyzing biological samples.

To obtain a comprehensive view of microbial communities, it is preferable to have larger inserts up to 200 kb, but obtaining high molecular weight DNA with sufficient purity is very challenging. Thus, we are developing new technologies for recovering extremely high molecular weight DNA from environmental samples. Our results showed that more than 200 kb fragments can be isolated and cloned.

The genomic technologies we have developed are important not only to environmental studies but also to the general field of microorganism detection and characterization. Our technology innovations will greatly advance researcher¡¯s capabilities to analyze microbial communities in the environment. These technologies will be important in addressing microbial problems associated with pathogen detection, microbial ecology of infectious diseases, plant growth, animal health, biodiversity, pharmaceutical discovery, bioprocessing of industrial products, waste-water treatment, and bioremediation of contaminants, because microbial communities are important in each of these areas.

Development of new methods for measuring codon usage bias

Many methods for measuring synonymous codon usage bias require standard codon reference. However, it is difficult to determine the standard codon reference when little biological information is available from an organism of interest. Thus, a novel information theory-based method for measuring synonymous codon usage bias was developed. This method does not require standard codon reference. Analysis showed that this method is advantageous over other existing methods.

Development of novel approach for network analysis

Although genomic technologies such as microarrays provide powerful tools for identifying cellular interaction networks consisting many individual modules, defining such modules without ambiguity is very difficult because all current methods rely on arbitrarily chosen thresholds and hence the results are subjective. Thus we have developed a novel, reliable, sensitive and robust approach for automatically identifying functional modules using a mathematically defined threshold predicted by random matrix theory (RMT). Applying this approach to microarray data from yeast, human, E. coli and S. oneidensis demonstrated that it correctly identifies functional modules with the expected properties consistent with general network theory. Experimental validation on the predicted functions of 10 unknown genes from yeast and Shewanella indicated that this approach is useful for predicting the functions of unknown genes. This approach will be ideal for analyzing high throughput genomics data for modular network identification and gene function prediction.
¡¡