2011 Annual Report
1a.Objectives (from AD-416)
To develop and validate a universal plant virus microarray for detection and differentiation of plant viruses. To demonstrate the ability of an oligonucleotide microarray to detect and differentiate plant viruses from random amplification of plant total nucleic acid extracts.
1b.Approach (from AD-416)
ARS will acquire the lists of viral taxa to be represented on the viral detection microarray, and virus-infected samples from which to amplify nucleic acids to validate the microarray. This information and material will be utilized by both ARS and the Cooperator to jointly develop and validate the microarray for detection of target viruses, and to make validation results available to collaborators via a web server. The COOPERATOR will further develop bioinformatic software (based on E-Predict and vTaxI) to perform analysis of viral sequences to identify suitable sequences for development of oligonucleotides, and for analysis of microarray hybridization results to determine with a high degree of confidence which viruses were present in validation samples.
Research efforts have focused on updating the Uchip software program for analysis of Universal Plant Virus Microarray (UPVM) results. More than 60 UPVM hybridizations have been analyzed using the Uchip program. Uchip allows users to organize microarray experiments into different studies and share data within these studies among users at different sites. UPVM users at three sites (University of Oklahoma, USDA-ARS Beltsville, MD, and the Danforth Center) have begun to use Uchip and have organized their hybridizations into separate studies.
The Uchip program allows users to compare their results to different sets of viral sequences and to use several noise models, which collaborators have done by running 414 different T-Predict calculations on UPVM arrays. Uchip has undergone continued improvement and development. The focus of this work has been improvement of the speed of analysis. Two changes in the architecture of the application have led to the greatest results. First, the T-Predict calculations were moved to servers at Utah, reducing the computational load on the remote site computers, but requiring that client computers be connected to the internet. The other major change was a redesign of the part of the database which stores the microarray intensities. In the field it was discovered that typical client computers using the database libraries operated faster when the intensities were partitioned into sub-tables with three or fewer arrays each. This "triplex design" prevents a profound slowdown when more that 50 arrays have been uploaded at a remote site. The University of Utah continues to provide technical support for Uchip users, including the three UPMV sites mentioned above. Those users also have provided, and continue to provide, critical feedback regarding the program functionality. A support website is available for users: http://fischer-lab.path.utah.edu/software/uchip. T-Predict iteration will be added to Uchip before the end of the year. UPMV users have now collected sufficient data to tune the analysis by weighting consistently noisy microarray probes such that they do not lead to spurious calls. It typically takes at least 50 microarrays to characterize these ‘bad’ probes on a novel platform, and a University of Utah scientist is now assembling these data and calculating the necessary weight values.
Communications to monitor progress were carried out by e-mail and conference calls between the various partners, and by written and oral reports to the NRI Plant Biosecurity Program.