Source: TEXAS A&M UNIVERSITY submitted to
A HOST-PATHOGEN APPROACH TO GWAS FOR ENHANCED RESISTANCE TO BACTERIAL MASTITIS IN DAIRY CATTLE
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
TERMINATED
Funding Source
Reporting Frequency
Annual
Accession No.
1008944
Grant No.
2016-67015-24923
Project No.
TEX09643
Proposal No.
2015-07033
Multistate No.
(N/A)
Program Code
A1221
Project Start Date
Feb 15, 2016
Project End Date
Feb 14, 2022
Grant Year
2016
Project Director
Seabury, C. M.
Recipient Organization
TEXAS A&M UNIVERSITY
750 AGRONOMY RD STE 2701
COLLEGE STATION,TX 77843-0001
Performing Department
Vet Med Pathobiology
Non Technical Summary
The primary goal of this project is to identify naturally occurring genetic variation that is either statistically predictive and/or causal for improved resistance to clinical mastitis in commercial U.S. dairy cows. Clinical mastitis is characterized by abnormalities in the milk or udder of dairy cows, which negatively impacts dairy production, thereby triggering economic losses, and diminished food security. Herein, we focus on clinical mastitis that is caused by Escherichia coli, with E. coli often being the most commonly isolated bacterial organism observed among U.S. dairy cows that experience single or recurring episodes of clinical mastitis. Relevant economic losses in the U.S. have been reported to exceed 700 pounds of saleable milk for a single cow experiencing clinical mastitis, which would include direct production losses, or the withholding of drug contaminated milk (i.e., therapeutic treatment) from the market. To reduce national dependence on therapeutic agents (i.e., antibiotics, anti-inflammatory drugs) while also enhancing dairy production, we take a novel approach which seeks to define natural genetic variation among U.S. production dairy cows and their E. coli clinical mastitis isolates that either jointly or independently enhance the risk of disease. At present, genetic data are already heavily utilized for accurately predicting many dairy production traits in the U.S., and form the basis for modern selective breeding programs (i.e., genomic selection). Information regarding naturally occurring genetic variants among U.S. dairy cows and E. coli bacterial isolates that enhance risk for clinical mastitis will be incorporated into existing U.S. dairy prediction and selective breeding programs. Selective breeding of more resistant U.S. dairy cows is ultimately the most sustainable approach to reducing disease incidence and associated production losses, as these cows will possess superior production efficiency, require less handling, experience improved animal welfare, require decreased use of therapeutic agents (i.e., antibiotics, anti-inflammatory drugs) and are ultimately the most profitable.
Animal Health Component
100%
Research Effort Categories
Basic
50%
Applied
50%
Developmental
(N/A)
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
30334101080100%
Knowledge Area
303 - Genetic Improvement of Animals;

Subject Of Investigation
3410 - Dairy cattle, live animal;

Field Of Science
1080 - Genetics;
Goals / Objectives
The primary research objective of this project is to identify bovine genetic variation associated with enhanced risk of E. coli CM in U.S. Holstein dairy cattle, and to use this information to foster the deployment of SNP-based selection tools that identify dairy cattle with reduced susceptibility. To increase the precision and likelihood of detecting associations with susceptibility, we will focus on the lifetime clinical mastitis distribution for a single bacterial pathogen (E. coli), thus creating a very precise animal disease phenotype. Moreover, genome sequencing data generated from the diagnostic isolates that define each E. coli CM case will also be used in a mixed-model approach to estimate the joint effects and/or interactions related to both host and pathogen genomic variation with respect to risk for recurrent E. coli CM in dairy cattle. Completion of these objectives will positively augment modern on-farm prevention and control measures, and therefore, is expected to allow for a significant reduction in recurrent E. coli CM among dairy cattle. To achieve our overarching primary research objective, we will employ the following specific objectives:Use > 9,000 precisely documented Holstein dairy cows from commercial NY dairies to define the extremes (i.e. tails) of the lifetime E. coli clinical mastitis distribution andselect cows from each extreme (i.e. multiple E. coli CM cases and controls with zero lifetime CM). To increase our sample size, we will also use these data to match single E. coli CM case cows and zero lifetime CM controls by lactation, age, and herd.Holstein cows with one or more (i.e. recurrent) episodes of diagnostically confirmed E. coli CM will be selected as "cases", with enrollment priority given to cows with ≥ 2 confirmed E. coli CM episodes. A CM severity score will also be assigned to case cows (see approach). Precisely monitored cows matched by lactation, age, and herd that display no lifetime evidence of CM will be selected as "controls." The ratio of controls to cases will be > 1:1 (n = 500 controls and 300 cases), which is determined by the prevalence of single and repeat E. coli CM cases prior to and across the award period.E. coligenomes representing the diagnostic isolates recovered from CM case cows will be sequenced to ≥ 40X average coverage using standard paired-end libraries applied to an Illumina HiSeq 2500 to facilitate E. coli strain classification, and to elucidate genome-wide information regarding strain-specific variation.All cows will be genotyped using the Illumina GGP-HD 135K assay and imputed to 778K density. Bovine 778K genotypes and E. coli variation will be bioinformatically joined for use in a novel genome-wide association analysis (GBLUP; EMMAX; Bayesian Models) that accounts for E. coli strain and/or polymorphism information with respect to risk for E. coli CM episodes and clinical severity (see approach) while also correcting for any pedigree-based stratification among the cows.Bovine and/or E. coli markers which are statistically predictive for risk of E. coli CM in Holstein dairy cows will be transferred to the stakeholders in a format that is cost-effective, efficient, and readily incorporated within other ongoing efforts to employ genomic selection in beef and dairy cattle (BRD-CAP Project; Bovine Feed Efficiency Project; Heifer Fertility Project; Bovine Functional Assay Project; Dairy Fertility Project).
Project Methods
The primary goal of this study is to identify genetic variation that is statistically predictive and/or causal for improved resistance to E. coli clinical mastitis in dairy cattle while also leveraging funded education programs to train the next generation of scientists in cutting-edge agricultural genomics and bioinformatics, with genomic selection technologies transferred to the stakeholders via publication and industry partnerships. To accomplish this, we hypothesize that a genome-wide association approach with precise, pathogen-specific mastitis phenotypes and knowledge of host (bovine) and bacterial genome-wide variants will result in genetic improvements that can be incorporated within genomic selection models already underway in dairy cattle. Moreover, we propose a novel host-pathogen approach that accounts for the effects of host (bovine) and pathogen (E. coli) genome-wide variation to reduce economic losses associated with recurrent episodes of clinical mastitis caused by the most commonly isolated bacterial organism (E. coli).Because host (bovine) and pathogen (E. coli) genome-wide variation will be included in our association study, we fully expect to find variants that are strongly associated with differential susceptibility to E. coli clinical mastitis, and will transfer this information to the stakeholders in a cost-effective format. Finally, we also provide opportunities for discovering E. coli strains that are significantly overrepresented among clinical mastitis case-cows, as well as the E. coli additive genetic merits that effect host (bovine) risk for clinical mastitis.

Progress 02/15/16 to 02/14/22

Outputs
Target Audience:The targeted audience during the current reporting period was the U.S. Holstein commercial dairy industry at large; including U.S. commercial dairy employees, scientists, veterinarians, and facility owners (i.e., general stakeholders); with a particular focus on common (recurring) animal health problems that reduce productivity (i.e., clinical mastitis) and profitability. In collaboration with QMPS at Cornell University, we continued to communicate project goals and needs to the original commercial dairy in Cayuga County NY, where our study was initiated, but also communicated with asecond commercial dairy and study site within the same county, for the purpose of bolstering enrollment and disseminating project information; with project goals and needs communicated to this dairy as well. Both commercial dairy cooperators were provided with ongoing description(s) of project requirements, goals/objectives, ongoing progress and how the project may ultimately benefit them via QMPS staff and our Cornell University collaborators; including the possiblity of genomic predictions and the production of genomically estimated breeding values. We also targeted graduate and undergraduate students by bringing research into teaching at the graduate and undergraduate level as well as during postdoctoral training. Changes/Problems:Because 153 E. colidiagnostic isolates for Holstein case cows were accidentally discarded, and because these cows had already been genotyped on the Illumina BovineHD 778K assay, an economic problem arose. Essentially we had to use funds originally budgeted for graduate student training to cover the costs associated with running the Illumina BovineHD 788K assay on newly sampled case cows (i.e, to obtainpaired E. coli diagnostic isolates) that we enrolled to replace those accidentally discarded at QMPS. We also determined that some of the diagnostic isolates that we received from QMPS were for case cows that we did not receive blood from, and as such, we could not create paired genomic data (host+pathogen). This caused a reduction in the anticipated sample size for Holstein E. coli case cows with paired E. coli genomic data (i.e., from sequencing the diagnostic isolates). Therefore, we did not have an adequate sample size to jointly fit host and pathogen data within a novel mixed model genome-wide association analysis, as originally planned. Instead, we fit models with the number of E. coli CM episodes in Holstein case cows as the dependent variable, and E. coli genomic features (i.e., known virulence loci) as fixed effects (i.e., the predictors). Nevertheless, despite this modification, we were able to detect significant associations between the presence of known E. coli virulence factors, and variation in the number ofE. coli CM episodes amongcommercial U.S. Holsteins. What opportunities for training and professional development has the project provided?Despite a rebudgeting that occurred to offset the loss of 153 diagnostic isolates (which entailed funding for graduate support), the project again provided research associate (Eric Bhattarai) and postdoctoral researchers (Dr Gizem Levent; Dr. Natalie Halbert)with some relevant laboratory or analystical training; in the form of DNA isolation (Bhattarai), DNA quantitation (Bhattarai), DNA archiving (Bhattarai),database managment (Bhattarai, Levent, Halbert), and various analyses (Levent, Halbert). Funding for these personel primarily came from other sources including indirect cost return; but the project itself provided for valuable cross training.Dr. Levent performed many of the primary analyses related to the genomic characterization of theE. coligenome sequencing data; which provided a significant training opportunity, but was also highly synergistic to the project goals and objectives. Dr. Seabury was also able to engage in professional development related to microbial genomics, and the analysis of microbial genomes; despite his primary role in the analysis of all bovine genomic data. Dr. Seabury also performed a variety of analyses which sought to evaluate the statistical relationship between the presence or absence of specific E. coli genomic features, and the number of E. coli CM episodes in commercial U.S.Holstein case cows. How have the results been disseminated to communities of interest?The results of this project have only been completed in the current reporting year; and close to the end date of the reporting period, because of a resampling initiative that was put into motion after 153 E. coli diagnostic isolates were accidentally discarded at QMPS (Cornell University). However, the PI will submit two open access manuscripts in September of 2022 for peer review and publication; with these manuscripts focusing on genomic relationship matrix heritability estimates, genome-wide association, and genomic prediction with cross validation for the binary case-control phenotype, and the total number of E. coli CM episodes in commercial U.S. Holsteins; the latter of which will also include the E. coli genomic analyses as well. We also plan to enage with animal health companies and commercial dairies who might be interested in deploying genomic prediction equations for E. coli CM in U.S. Holsteins. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? During the current reporting period, we completed enrollment of 334 E. coli clinical mastitis (CM) Holstein case cows, and 505 Holstein control cows (Objectives 1-2). Among the study cows, the number of diagnostically confirmed E. coli CM events ranged from 0 to 5; with parity ranging from 1 to 11 (mean parity = 4). Thus, two animal disease phenotypes were available (i.e., binary case-control; total episodes of diagnostically confirmed E. coli CM) and used to produce genomic relationship matrix (GRM) heritability estimates, genome-wide association analyses (GWAA), and genomic predictions with cross validation. GRM heritability estimates for the binary case-control phenotype ranged from 0.16± 0.07 to 0.20± 0.09 using EMMA variance component analysis and/or GCTA to produce estimates on the liability scale (i.e., with weighted mean prevalence). Using the same methods, GRM heritability estimatesfor the number of diagnostically confirmedE. coli CM episodes were uniformly lower.GWAA performed using single and multi-marker mixed linear models revealed evidence for 8or more QTL located on 6bovine chromosomes (i.e., BTA2, BTA5, BTA13, BTA17, BTA18, BTA29) which met a nominal significance threshold for polygenic effects (P≤ 5e-05). Several positional candidate genes underlying the detected QTL were biologically relevant to E. coli infections in multiple mammalian species.Mean genomic prediction accuracies (binary case-control phenotype) using genomic best linear unbiased prediction (GBLUP) in conjunction with k-fold cross validation (i.e.,k= 3; 10 replicates) were moderate (0.63), but higher than those recently estimated for 9 economically importanthealth traits in commercial U.S. Holsteins (Anim Genet. 2020 Mar; 51(2): 192-199); therebyindicating thatgenomic predictions for pathogen-specific E. coli CM may be a useful addition to moderngenomic selection initiatives already underway in U.S. Holstein dairy cattle. To investigate whether E. coli genomic features (i.e., genetic variation) influenced recurring episodes of E. coli CM in U.S. Holstein dairy cows, we fit several models using multiple methodswhereby E.coli virulence loci were treated asfixed effects (i.e., predictors);with the total number of E. coli CM episodesdesignated as the dependentvariable. At least two known E. coli virulence locidisplayed evidence of significant associations with variation in E. coli CM episodes among commercial U.S. Holstein case cows.Given the accidental discarding of 153 E. coli diagnostic isolates for commercial U.S. Holstein case cows at QMPS (Cornell University), our ability to jointly fit one or more mixed linear models with host and pathogen genomic data as originally planned was not possible; due to sample size limitations. Nevertheless, all experimental objectives (1-4) were completed; with one modification to the host-pathogen analysis originally planned, and two manuscripts (objective 5) will provide the project dissemination necessary for stakeholder awareness, and for genetic improvement related to E. coli CM in commercial U.S. Holsteins.

Publications


    Progress 02/15/20 to 02/14/21

    Outputs
    Target Audience:The targeted audience during the current reporting period wasthe U.S. Holstein commercial dairy industry at large; including U.S. commercial dairy employees, scientists, veterinarians, and facility owners (i.e., general stakeholders); with a particular focus on common (recurring) animal health problems that reduce productivity (i.e., clinical mastitis) and profitability. In collaboration with QMPS at Cornell University, we continued to communicate project goals and needs to the original commercial dairy in Cayuga County NY, where our study was initiated, but also enrolled a second commercial dairy in the same county, for the purpose of bolstering enrollment; with project goals and needs communicated to this dairy as well. Both commercial dairy cooperators were provided ongoing description of project requirements, goals/objectives, ongoing progress and how the project may ultimately benefit them via QMPS staff and our Cornell University Collaborators; including the possiblity of genomic predictions and the production of genomically estimated breeding values. We also targeted graduate and undergraduate students by bringing research into teaching via direct mentoring of graduate students and/or introduction to project related concepts during undergraduate education, and during postdoctoral training. Changes/Problems:Because 153 E. coli isolates for Holstein case cows were accidentally discarded, and because these cows had already been genotyped on the Illumina BovineHD 778K assay, an economic problem arose. Essentially we had to use funds originally budgeted for graduate student training to cover the costs associated with running the Illumina BovineHD 788K assay on the new case cows (with paired E. coli diagnostic isolates) that we enrolled to replace those accidentally discarded at QMPS. We also determined that some of the diagnostic isolates that we received from QMPS were for case cows that we did not receive blood from, and as such, we could not create paired genomic data (host+pathogen). What opportunities for training and professional development has the project provided?Despite a rebudgeting that occurred to offset the loss of 153 diagnostic isolates (which entailed funding for graduate support), the project again provided graduate student (Miranda Wilson) and research associate (Eric Bhattarai) with some relevant laboratory training; in the form of DNA isolation, DNA quantitation, DNA archiving, and database managment. Primary funding for these personell came from other sources and indirect cost return; but the project itself provided for valuable cross training for on graduate student. Likwise, two postdoctoral researchers were also exposed to project objectives and deliverables (Dr. Natalie Halbert; Dr. Gizem Levent); thus leveraging learning opportunities. Dr. Levent performed many of the primary analyses related to the genomic characterization of the E. coli genome sequencing data; which provided a significant training opportunity, but was also highly synergistic to the project goals and objectives. How have the results been disseminated to communities of interest?Our Cornell collaborators (Including QMPS Staff) maintained the dissemination of the importance of the study to both the participating commercial NY dairies; including how the deliverables could potentially help them reduce the prevalence of E. coli clinical mastitis via genomic prediction. The final results of this study come via the joint analysis of bovine and E. coli genomic data; and while the host analyses were completed in the current reporting period; the role of specificE. coli virulence factors on the risk forrecurrent episodes of E. coli clinical mastitis was still under investigation.We did also engage with several companies in the animal health industry;all of which provide services to the U.S. dairy industry at large. We did so to disseminate project goals and deliverables; in hopes of creating opportunities for leveraging, and disseminating the final project deliverables to the stakeholders. We expect to leverage private research partnerships with commercial dairy producers as a result of this project, and have generated a project abstract for CRWAD 2023. What do you plan to do during the next reporting period to accomplish the goals?We plan to finalize all analyses and results, with the drafting at least one manuscript for peer-reviewed publication, to help disseminate project results and deliverables.

    Impacts
    What was accomplished under these goals? During the current reporting period of the study, we continued to collect the required samples, as described in Objectives 1 and 2; with identification and enrollment of 141Holstein controls and 134Holstein E. coli clinical mastitis case cows. The E. coli diagnostic isolates for all Holstein case cows were retained and archived at -80C for use in Objective 3 (genome sequencing), with bovine blood obtained from all cows for DNA isolation at Texas A&M. Herein, Objective 3 (E. coli genome sequencing) was executed using an Illumina novaseq, thus delivering 40X-50X theoretical genome coverage for all E. coli diagnostic isolates obtained to date (n = 169); excluding those that were accidentally discarded by QMPS at Cornell University (n = 153). Likewise, all study cows were genotyped(Objective 4; multiples of 96) in using the Illumina BovineHDSNP array; thus resulting in the production of high density genotypes for 839 cows (n = 334 E. coli clinical mastitis (CM) cases; n = 505 controls); not including those that failed call rate quality control metrics. Genome-wide association with genomic relationship matrix heritabilty estimates were performed (single and multi-marker approaches), and the accuracy of genomic predictions were assessed using cross validation. Genome sequencing data for all E. coli diagnostic isolates provided by Cornell (QMPS) were analyzed using resources inside and outside of the Center for Genomic Epidemiology (Lead by Postdoc Gizem Levent PhD); thus resulting in MLST genotypes, plasmid characterization, and the genomic characterization of antimicrobial resistance genes as well as virulence factors. Statistical models were fit using E. coli genomic features, and the number of recurring E. coli CMevents among the case cows. Therefore, progress in the current reporting periodpertained to Objectives 1-5.

    Publications


      Progress 02/15/19 to 02/14/20

      Outputs
      Target Audience:The targeted audience during the current reporting period was the U.S. Holstein commercial dairy industry at large; including U.S. commercial dairy employees, scientists, veterinarians, and facility owners (i.e., general stakeholders); with a particular focus on common (recurring) animal health problems that reduce productivity (i.e., clinical mastitis) and profitability. In collaboration with QMPS at Cornell University, we continued to communicate project goals and needs to the original commercial dairy in Cayuga County NY, where our study was initiated, but also enrolled a second commercial dairy in the same county, for the purpose of bolstering enrollment; with project goals and needs communicated to this dairy as well.Both commercial dairy cooperators wereprovided ongoingdescription of projectrequirements, goals/objectives, ongoing progress and how the project may ultimately benefit them via QMPS staff and our Cornell University Collaborators; including the possiblity of genomic predictions and the production of genomically estimated breeding values.We also targeted graduate and undergraduate students by bringing research into teaching via direct mentoring of graduate students and/or introduction to project related concepts during undergraduate education. Changes/Problems:The only real problem encountered during this reporting period was the prior discovery that most of our E. coli diagnostic isolates for Holstein case cows were accidentally discarded at QMPS (Cornell University) during a previous reporting period; and notably,all of these case cows had already been genotyped on the IlluminaHD 778K assay, in anticipation of receiving the diagnostic isolates. Notably, all of the case cows can still be used in host genome wide association analysis, with genomic relationship matrix heritability estimates, and genomic predictions with cross validation, but the host-pathogen aspect of the analysis (joint effects) cannot ensue without paired data (i.e., case-cows with recurrent episodes of E. coli clinical mastitis + the E. coli diagnostic isolates that are characterized by genome sequencing).Therefore, we continue enrollment and sampling. What opportunities for training and professional development has the project provided?Despite a rebudgeting that occurred to offset the loss of 153 diagnostic isolates (which entailed funding for graduate support), the project again provided graduate student (Miranda Wilson) and research associate (Eric Bhattarai) with some relevant laboratory training; in the form of DNA isolation, DNA quantitation, DNA archiving, and database managment. Primary funding for these personell came from other sources and indirect cost return; but the project itself provided for valuable cross training for on graduate student. How have the results been disseminated to communities of interest?Our Cornell collaborators (Including QMPS Staff) maintained thedissemination of the importance of the study to both the participating commercial NY dairies; including how the deliverables could potentially help them reduce the prevalence of E. coliclinical mastitis via genomic prediction. However, because the final genomic data could not be generated during this reporting period, the direct impact(s) of the study to the participating NY dairy and the broader stakeholder groups could not yet be demonstrated (or even well discussed as the current sample size precluded confident scientific inferences). The final results of this study come via the joint analysis of bovine and E. coli genomic data; and none of those analyses can be finalized until the targeted sample size is reached. The reason for this is the quoted price on the cattle SNP array, which requires a specific sample size (multiples of 96), and the quoted price on the E. coli genome sequencing, which requires that all diagnostic isolates be multiplexed into a single lane of Illumina sequencing (as budgeted in the grant). A poster was also presented on the project at the USDA PD meeting. We did also engage with several companies in the animal health industry; all of which provide services to the U.S. dairy industry at large. We did so to disseminate project goals and deliverables; in hopes of creating opportunities for leveraging, and disseminating the final project deliverables to the stakeholders. What do you plan to do during the next reporting period to accomplish the goals?In the next reporting period, we primarily plan to continue working on enrollment via Objectives 1 and 2; in an effort to collect/recollect the targeted sample size, which we hope to achieve, so that E. coli and bovine genomic data can be generated for the primary analyses. In the initial 2 years of the study, we successfully obtained a small batch of E. colidiagnostic isolates from Holstein clinical mastitis case cows, and prepped those isolates for a proof of concept run that included: Successful robotic DNA isolation at Texas Veterinary Medical Diagnostic Laboratory (with collaborator Dr. Pam Ferro), Successful Illumina library prepration, Successful Illumina sequencing to approximately 100x, and Successful genomic characterization of each isolate via the Center for Genomic Epidemiology (CGE). Should we reach our targeted sample sizes, we will generate all remaining genomic data and perform the final analyses. However, given thenecessary restart on sampling, and the potential effects of COVID-19, we may need additional time to identify enough samples for a meaningful analysis.

      Impacts
      What was accomplished under these goals? During the current reporting period of the study, we continued to collectthe required samples, as described in Objectives 1 and 2; with identification and enrollment of 56Holstein controls and 30 Holstein E. coli clinical mastitis case cows. The E. coli diagnostic isolates for all Holstein case cows were retained and archived at -80C for use in Objective 3 (genome sequencing), with bovine blood obtained from all cows for DNA isolation at Texas A&M. Although we successfully obtained a small batch of E. coli diagnostic isolates for validation of our laboratory and bioinformatic procedures early on in the project, Objective 3 (E. coli genome sequencing) could not be initiated until all case cows and their diagnostic isolates were obtained; as budgeted in the grant (for multiplex in a single lane). The accidental discarding of 153 isolates at QMPS essentially caused a restart in the sampling of E. coli case cows with paired E. coli diagnostic isolates; thus two commercial dairies were used given the observed prevalence, to acclerate enrollment within the established project timelines. Likewise, all cows must be genotyped in large batches (Objective 4; multiples of 96) in order to receive the discount in the quoted price for the SNP array. Therefore, most progress pertained to Objectives 1 and 2 (sampling/enrollment).

      Publications


        Progress 02/15/18 to 02/14/19

        Outputs
        Target Audience:The targeted audience during the current reporting period was the U.S. Holstein commercial dairy industry at large; including U.S. commercial dairy employees, scientists, veterinarians, and facility owners (i.e., general stakeholders); with a particular focus on common (recurring) animal health problems that reduce productivity (i.e., clinical mastitis) and profitability. In collaboration with QMPS at Cornell University, we initiated our study at a commercial dairy in Cayuga County NY. The commercial dairy cooperator was provided some description of the projects requirements, goals/objectives, ongoing progress and how the project may ultimately benefit them via QMPS staff and our Cornell University Collaborators. We also targeted graduate and undergraduate students by bringing research into teaching via direct mentoring of graduate students and/or introduction to project related concepts in undergraduate education. Changes/Problems:The only real problem encountered during this reporting period was the discovery that most of our E. coli diagnostic isolates for Holstein case cows were accidentally discarded at QMPS (Cornell University). Specifically, 181 Holstein E. coli clinical mastitis (case) cows were identified, and their E. coli diagnostic isolates were archived at -80C. Moreover, we had recently generated Illumina 778K HD genotype data for all 181 Holstein E. coli case cows, and 304 Holstein control cows. Therefore, while the bovine 778K HD data is useful for heritabiltiy estimates, genome-wide association, and genomic prediction with cross validation, the ability to perform joint genome analyses (Host-Pathogen) using E. coli genome sequencing data and bovine genotype data was inhibited. Only 28 of the 181 isolates were available for genome sequencing. Therefore, we immediately identified another suitable commercial dairy study site nearby, and began following Holstein cows there as well; to bolster our ability to identify additional samples (especially E. coli case cows for which E. coli diagnostic isolates would be available). Recognizing that this would increase the cost of achieving the targeted sample size, the PD repurposed his budget line item for graduate student support, and used that money to cover costs associated with additional sampling. One main cost of additional sampling is the need to run more Illumina 778K HD assays than initially planned, because each E. coli case cow was supposed to have a paired E. coli diagnostic isolate. Nevertheless, we pressed forward, and were still able to engage graduate students (i.e., Miranda Wilson) using funds outside of this grant. Moreover, we recognized (given the prevalence of E. coli clinical mastitis) that we would need a project extension, and we planned for that accordingly. What opportunities for training and professional development has the project provided?The project again provided graduate student (Miranda Wilson) and research associate (Eric Bhattarai) with some relevant laboratory training; in the form of DNA isolation, DNA quantitation, DNA archiving, and database managment. In anticipation of receiving the necessary data to carry out Objectives 3-5, one graduate student was able to engage in cross-training for the purpose learning next generation sequencing analysis as well as advanced concepts related to genome-wide association. How have the results been disseminated to communities of interest?Our Cornell collaborators (Including QMPS Staff) took the lead on providing the initialdissemination ofthe importance of the study to the participating NY dairy; including how the deliverables could potentially help them reduce the prevalence of E. coli clinical mastitis via genomic prediction. However, because the final genomic data could not be generated during this reporting period, the direct impact(s) of the study to the participating NY dairy and the broader stakeholder groups could not yet be demonstrated. The final results of this study come via the joint analysis of bovine and E. coli genomic data; and none of those analyses can be finalized until the targeted sample size is reached. The reason for this is the quoted price on the cattle SNP array, which requires a specific sample size (multiples of 96), and the quoted price on the E. coli genome sequencing, which requires that all diagnostic isolates be multiplexed into a single lane of Illumina sequencing (as budgeted in the grant). A poster was also presented on the project at the USDA PD meeting. What do you plan to do during the next reporting period to accomplish the goals?In the next reporting period, we primarily plan to continue working on enrollment via Objectives 1 and 2; in an effort to collect/recollectthe targeted sample size, which we hope to achieve, so that E. coli and bovine genomic data can be generated for the primary analyses. In the initial 2 years of the study, we successfully obtained a small batch of E. coli diagnostic isolates from Holstein clinical mastitis case cows, and prepped those isolates for a proof of concept run that included: Successful robotic DNA isolation at Texas Veterinary Medical Diagnostic Laboratory (with collaborator Dr. Pam Ferro), Successful Illumina library prepration, Successful Illumina sequencing to approximately 100x, and Successful genomic characterization of each isolate via the Center for Genomic Epidemiology (CGE). Should we reach our targeted sample sizes, we will generate all remaining genomic data and perform the finalanalyses.

        Impacts
        What was accomplished under these goals? During the current reporting period of the study, we continued tomakestrides toward collecting the required samples, as described in Objectives 1 and 2; with identification and enrollment of 132Holstein controls and 76 Holstein E. coli clinical mastitis case cows. The E. coli diagnostic isolates for all Holstein case cows were retainedand archived at -80C for use in Objective 3 (genome sequencing), with bovine blood obtained from all cows for DNA isolation at Texas A&M. Although we successfully obtained a small batch of E. coli diagnostic isolates for validation of our laboratory and bioinformatic procedures early on in the project, Objective 3 (E. coli genome sequencing) could not be initiated until all case cows and their diagnostic isolates were obtained; as budgeted in the grant (for multiplex in a single lane). Likewise, all cows must be genotyped in large batches(Objective 4; multiples of 96) in order to receive the discount in the quoted price for the SNP array. Therefore, most progress pertained to Objectives 1 and 2; although we did start genotyping the Holstein E. coli clinical mastitis case cows, and the Holtstein controls obtained to present on the Illumina 778K HD assay instead of the GGP135K or GGP150K assay. The rationale for this change was a renegotiation of price related to the Illumina 778K. Essentially, we could purchase this assay for the same price as the GGP assays; without the need to perform any imputation at all. Therefore, we utilized this opportunity to collect 778K SNP data for all Holstein cases and controls acquired to date, and perform initial genomic relationship matrix heritability estimates and genome wide association. Some controls did convert to cases across time, so we endeavored to obtain their E. coli diagnostic isolatesas well. Within this reporting period, having 181 E. coli case cows, we advised Cornell University that we were ready to take possession of all archived E. coli diagnostic isolates for case cows genotyped on the Illumina 778K HD assay; for the purpose of robotic DNA isolation and storage until such time that 300 Holstein case-cows with paired E. coli diagnostic isolates had been acquired for E.coli genome sequencing. To our teams dismay, we learned that 153 E. coli diagnostic isolates had been mistakenly discarded at QMPS during a time where staff changes had occurred. Therefore, we endeavored to monitor, enroll and sample Holsteins with even greater vigor;and identified a group of stakeholder cooperators at a second commercial dairy that was willing to serve as a project study site. We immediately began to follow Holsteins at this second nearby commercial diary study site, to identify additional E. coli clinical mastitis case cows with paired E. coli diagnostic isolates as well as suitable Holstein control cows.

        Publications


          Progress 02/15/17 to 02/14/18

          Outputs
          Target Audience:The targeted audience during the current reporting period was the U.S. Holstein commercial dairy industry at large; including U.S. commercial dairy employees, scientists, veterinarians, and facility owners (i.e., general stakeholders); with a particular focus oncommon (recurring)animal health problems that reduce productivity (i.e., clinical mastitis) and profitability. In collaboration with QMPS at Cornell University, we initiated our study at a commercial dairy in Cayuga County NY. At the time, QMPS was already interacting with this dairy, and notification of the funding for this study allowed our team to hit the ground running; which even included some banked samples used as preliminary data for the original grant application. The targeted audience was provided some description of the projects requirements, goals/objectives, and how the project may ultimately benefit them via QMPS staff and our Cornell University Collaborators. Changes/Problems:No major changes or problems were encountered during this reporting period. What opportunities for training and professional development has the project provided?The project provided graduate student (Miranda Wilson) and research associate (Eric Bhattarai)training; in the form of DNA isolation, DNA quantitation, DNA archiving, anddatabase managment. In anticipation of receiving the necessary data to carry out Objectives 3-5, one graduate student was able to engage in cross-training for the purpose learning next generation sequencing analysis as well as advanced concepts related to genome-wide association. How have the results been disseminated to communities of interest?Our Cornell collaborators (Including QMPS Staff) took the lead on disseminated the importance of the study to the participating NY dairy; including how the deliverables could potentially help them reduce the prevalence of E. coli clinical mastitis via genomic prediction. However, because no genomic data could be generated during this reporting period, the direct impact(s) of the study to the participating NY dairy could not yet be demonstrated. What do you plan to do during the next reporting period to accomplish the goals?In the next reporting period, we primarily plan to continue working on enrollment via Objectives 1 and 2; in an effort to collect the targeted sample size, so that E. coli and bovine genomic data can be collected for the primary analyses.

          Impacts
          What was accomplished under these goals? During the first year of the study, we made great strides toward collecting the required samples, as described in Objectives 1 and 2; with identification and enrollment of172 Holstein controls and 105 HolsteinE. coli clinical mastitiscase cows. The E. coli diagnostic isolates for all Holstein case cows were available and archived at -80C for use in Objective 3 (genome sequencing), withbovine blood obtained from all cows for DNA isolation at Texas A&M. Objective 3 (E. coli genome sequencing) could not be initiated until all case cows and their diagnostic isolates were obtained; as budgeted (for multiplex in a single lane). Likewise, all cows must be genotyped together (Objective 4) in order to receive the discount in the quoted price for the SNP array. Therefore, all progress pertained to Objectives 1 and 2.

          Publications


            Progress 02/15/16 to 02/14/17

            Outputs
            Target Audience:The primary audience targeted by this project, in all reporting periods, includes the stakeholders (private dairy operations) and other scientists. Since the start of the project we have engaged two private New York commercial dairies for sampling and cooperation, along with diagnosticians, veterinarians, and extension associatesat the Texas Veterinary Medical Diagnostic Laboratory (Texas A&M University), the Quality Milk Production Services-Animal Health Diagnostic Center (Cornell University), and a large group of private southern dairy producers during a BRD-CAPcontinuing educationpresentation in the Texas Panhandle (June 18, 2016), wherebythis project was introduced. We have also targeted graduate and veterinary students at Texas A&M University, and Cornell University, for training related to carrying out the primary objectives of this project. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?As the PD of this project, I have begun to expose and mentortwo graduate studentsin relation tothe basic hypotheses and statistical approaches that will be needed to carry out and complete the proposed objectives. I have also taken advantage of a continuing education program on BRD (as presenter, June 18, 2016) to introduce this project to our southern dairy producers (Texas Panhandle, and New Mexico).Undergraduate students currently working in the laboratory are also benefiting from being mentored by graduate students that will be analytically involved in this project. How have the results been disseminated to communities of interest?Throughout the first year of the study, and at present, we are still in a data collection phase of the project, and have not yet generated the proposed results. Therefore, we are only disseminating an introduction to the project at the moment; to both the stakeholder and scientific communities. What do you plan to do during the next reporting period to accomplish the goals?Across the next reporting period, we plan to continue our longitudinal observations and sampling, especially in relation to E. coli case cows. At present, we have some archived samples from case cows that can be used as well ascase cows recently identified for sampling, and/or those that have recently been sampled. We want to achieve our targeted number of cases and controls, which is precisely why weproposed afour year award period. However, we do expect to start sequencing and analyzing E. coli genomes during the next reporting period, and to perhaps start genotyping the dairy cattle with a medium density SNP array, as originally proposed.

            Impacts
            What was accomplished under these goals? For the specific objectivesproposed in this study, objectives one and two have been partially fulfilled during the first calendar year. This includes sampling of 426 total Holstein dairy cows, from two private New York dairies. Among those sampled, 27 were suitable for enrollment ascases (i.e., 1 or more confirmed case ofE. coliclinical mastitis), and 390 were suitable for tentative enrollment as controls, althoughwe will follow them longitudinally for verification until all of the cases have been enrolled. Notably, we sampled vigorously during the first year to ensure enrollment of older Holstein dairy cows with appropriate phenotypes, because some of these cows are nearing their production lifetimes, while others can be followed for the duration of the award period. Genomic DNA was isolated from the whole blood of all426 Holstein dairy cows sampled, and subsequently banked in a -80C repository for genotyping by SNP array. Likewise, we also isolated DNA from the E. coli diagnostic isolatesobtained fromall case cows, and banked those DNAs in a similar -80C repository for genome sequencing. Importantly, where possible, we actually isolated DNA fromseveralE. coli colonies recovered from each case cow, to enable future studies beyond the objectives required to complete this project. We have not yet commenced with genome sequencing of the E. coli diagnostic isolates because we are waiting for a larger batch, to take advantage of the economy of scale afforded byour high-yielding Illumina sequencers. Likewise, we also have not yet started to genotype the Holstein dairy cows via SNP array, becasue again, we are waiting for a larger batch, to take advantage of discounts related to large-sample pricing, and the potential for price reduction(s). Our quote gaurantees the price budgeted for this project, but costs may decline, particularly for higher density SNP arrays, and we want to be in a position to take advantage of that should it occur.

            Publications