Progress 01/01/13 to 09/30/13
Outputs
Target Audience:Our primary target audiences consist of dairy producers, dairy veterinarians, undergraduate/graduate students, and the stakeholders at large. However, we also use data and analyses for innovative teaching methods in two undergraduate courses. These activities simply entail "bringing research into teaching", with lessons that are either hands-on (analysis) or scenario-based (practical applications of the research). Our efforts this reporting period have reached both undergraduate and graduate students. We have not yet formally disseminated information to producers, dairy veterinarians, and the stakeholders at large because we wish to first publish the results in a peer-reviewed journal. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?This project provided one undergraduate student (Aaron Bonnett), one graduate student (Yvette Halley), and one research associate (Eric Bhattarai M.S.) with detailed training (data management, compilation, and association tests). Mr Bonnett and Bhattarai received the most detailed hands-on instruction with the Project Leader Dr. Seabury. Yvette Halley also received additional hands on training. We anticipate a publication in 2015. How have the results been disseminated to communities of interest?The results and methods have been disseminated to undergraduate and graduate students, as a vehicle for teaching the analytical methods and practical applications of the research. Beyond this, we have not yet disseminated the research to our target audiences or communities of interest because we are still finishing the primary analysis for publication in a peer-reviewed journal. Efforts to bolster our sample size are currently underway, to determine if SNP effects are uniform (directionally) among additional populations. What do you plan to do during the next reporting period to accomplish the goals?We plan to evaluate haplotype effects for inclusion into the final manuscript. The primary reason for a delay in the final manuscript centers squarely around attempting to determine whether or not a few SNPs that seem to potentially elicit small to moderate effects on MAP shedding can be validated in additional samples (i.e., increased sample size).
Impacts
What was accomplished under these goals?
Major goals and objectives include: 1 Genotype Holstein cows (n = 68) from our previous study [1] for new SNPs discovered during TLR pyrosequencing of the NADC cows (n = 150). Note, these SNPs weren't known at the time of our previous study [1], and the Holstein cows in our previous study were simply genotyped for every known TLR variable site that had been enumerated by sequencing a bovine DNA discovery panel consisting of > 31 breeds [1-2]. Therefore, adding these additional data into the analysis may drive the SNP effects toward increasing levels of significance, by simply increasing the sample size. Status: Objective 1 has been accomplished. 2 Some of the new TLR SNPs discovered by pyrosequencing of the NADC cows are within compositionally difficult regions of the targeted genes. This includes regions containing homopolymers. Therefore, a subset of cows produced ambiguous "intermediate" genotypes for a few of these markers (Figure 3), and therefore, they must be rerun to essentially fill in for "missing" data. This is normal, and we often will tweak the conditions and/or the SNP assay primers themselves to increase optimization related to difficult templates. Adding these missing data into the analysis may drive the SNP effects toward increasing levels of significance, by simply increasing the sample size. Status: Objective 2 has been accomplished 3 By accomplishing objective 1 above, we anticipate that some missing genotypes will occur for the same or similar reasons in regards to the 68 Holstein cows included in our previous study [1], and therefore, anticipate some reruns associated with the genotyping assays. Status: Objective 3 has been accomplished 4 Once the complete and final data are assembled from objectives 1-3, we can then repeat the analysis presented in Figures 1 and 2. Moreover, we also need to determine the best and most meaningful way to analyze the data. For example, once the final complete set of genotypes are assembled for the 218 Holstein cows, we can also perform a categorical regression, which would seek to essentially bin the shedding phenotypes (i.e. CFUs) into categories such as Low, Medium, High/Super Shedder, etc. Some precedence for this classification does exist within the literature [3-4], but we cannot construct meaningful categories until a complete final dataset has been produced for all cows, and the full distribution of CFUs can be plotted and visualized. Categorical regression and increased sample size may drive the SNP effects toward increasing levels of significance. Status: Objective 4 has been recently accomplished, but we are still evaluating different methods for stratifying the shedding phenotypes before publication, and we are also evaluating one additional analytical method for estimating SNP effects. At present, we are attempting to acquire additional samples for the purpose of investigating two SNPs that may elicit small to moderate effects on MAP shedding.
Publications
|
Progress 01/01/12 to 12/31/12
Outputs
OUTPUTS: A total of 745,652 Roche 454 pyrosequencing reads spanning the bovine TLRs were generated for Holstein cattle naturally infected with MAP using a multiplexed amplicon library approach previously described by PD Seabury. SNPs were predicted from the pyrosequencing reads using published procedures, and all new SNPs were incorporated into the existing custom SNP assays previously described by PD Seabury's laboratory. PD Seabury trained two graduate students and one Research Associate in Bioinformatics and Computational Genomics with this grant opportunity, and will publish the results when finished. Evidence for turning previous FAH funding into publications and training opportunities can be found in Ms Colleen Fisher's thesis and PLoS ONE paper. Similarly, we intend to publish our findings here as well. With respect to our project goals and workflow: 1) All of the DNAs were standardized to an appropriate working concentration (50ng/μl); 2) All variants (SNPs and indels) from TLR loci recognizing bacterial ligands were genotyped across our MAP-infected cattle via fluorescent genotyping assay; 3) A regression based on CFU phenotypes (High-to-Super shedders vs Low Shedders) was performed on 150 of the 218 cattle thus far (68%), and was completed for the purpose of progress report. This analysis is essentially a case-control study, with either categorial categorical CFU phenotypes, or colony numbers used as a semi-quantitative trait. We are also in the process of evaluating other modern analysis methods that may potentially be more powerful including: 1) Linear Mixed Models; and 2) Genomic Best Linear Unbiased Prediction (GBLUP). PD Seabury is currently concatenating data from two sources to facilitate the final 218 animal analysis. PARTICIPANTS: One MS graduate student (Ms Colleen Fisher) used this study and a related preceding study to earn a MS degree, and published her initial findings in PLoS ONE with Dr. Seabury. In addition to Ms Fisher, Dr. Seabury also trained Eric Bhattarai M.S. (RA) with respect to bioinformatics, SNP genotyping, SNP cluster calling, next generation sequence analysis,and is in the process of training him to perform regression and GBLUP. Likewise, Dr. Seabury is also utilizing this research project to train a Ph.D student in project-related bioinformatics, SNP genotyping, SNP cluster calling, next generation sequence analysis, regression, and GBLUP. Dr. Seabury actively leads his team by example. TARGET AUDIENCES: Nothing significant to report during this reporting period. PROJECT MODIFICATIONS: Nothing significant to report during this reporting period.
Impacts
All 150 Holstein Cattle naturally infected with MAP have been successfully genotyped for all known variable sites within the targeted bovine TLR genes. This includes new SNPs discovered during a series of pyrosequencing studies. An initial regression based on MAP fecal culture phenotypes has been performed, and includes correction for herd-related effects. This initial analysis identified some bovine TLR SNPs with uncorrected P-values of 0.00868 and 0.00761, respectively. Correction for multiple testing using the FDR approach to adjust the raw P-values (Benjamini and Hochberg; R code)produced corrected P-values of 0.0868 and 0.0868, respectively. We are currently adding data from another source to pool within a meta-analysis that includes cattle from 10 or more states. These initial results constitute a change in knowledge regarding bovine genetic variation that may potentially modulate small effects on MAP shedding in holstein dairy cattle.
Publications
- No publications reported this period
|