Local tests for local viruses

Project Leader: Dr Monica Kehoe, WA Department of Primary Industries and Regional Development (DPIRD)

Plant viruses cause diseases in food crops that result in losses in quality and production valued at over $30 billion a year worldwide.  Just like the common cold virus, plant viruses can evolve over time into different strains or new species, making their early detection and identification an ongoing challenge.  Dr Monica Kehoe at the WA Department of Primary Industries and Regional Development (DPIRD) is using next-generation sequencing methods and Pawsey supercomputing for data analysis to develop diagnostic tests that can keep up with local plant virus variations.

Core hours
mins potential for virus detection
week for genome results instead 18 mths
Partner Institution: WA Department of Primary Industries and Regional Development (DPIRD) System: Magnus, Zeus Areas of science: Plant Pathology Applications used: Nextflow, Singularity and various Bioinformatics software containers.

The Challenge

Grapevine leaf-roll viruses (GLRV) infect grape vines and are a major problem for grape growers across Australia, causing poor fruit quality and reducing yield between 10 and 70 per cent.  There are at least four different GLRVs, and more are being identified each year in grape growing countries. These viruses are typically transmitted between plants by mealy bugs.  However, each different GLRV can produce a range of different symptoms and severities.  Some plants may be asymptomatic but contribute to the spread of the virus in a vineyard, while other plants may display similar symptoms but be suffering from poor environmental conditions, like limited nutrition.  In addition, symptom expression can vary between different varieties of grape.

Unlike the common cold in people, once a plant becomes infected by a virus, it has it for life.  Therefore, growers can only contain the virus by managing its spread or removing and destroying all infected plants.  Early detection of virus infection then becomes critical when considering the health of a block of 40-year old shiraz vines, or when setting up a brand-new block of chardonnay vines.

Unfortunately, the available diagnostic tests for GLRVs are mostly produced overseas, and are a bit ‘hit and miss’, according to Australian growers.

“The genetic diversity of GLRV 1 and 3, the two major viruses of concern in Australia, is largely unknown, especially in Australia,” explains Dr Kehoe, a plant virologist and molecular plant pathologist for WA DPIRD.  “There are 56 complete genomes available in the public database for GLRV 3, but only one is Australian.  And there are only seven genomes of GLRV 1 available, none from Australia.  This is potentially a problem because we don’t know if the primers used for our current tests, which are based on those known genetic sequences, will pick up all of the GLRV1 and GLRV3 strains circulating in our vineyards.  Also, if strains emerge here that are different, those tests could fail.”

Dr Kehoe is using next-generation genetic sequencing technologies to develop new diagnosis methods and faster tests for GLRV 1 and 3.  She collected 264 samples from vineyards across south-west Western Australia for analysis, 103 testing positive for GLRV 1, and 123 positive for GLRV 3.  The next step was to sequence the genomes of these Australian virus isolates.

The size of that task shouldn’t be underestimated, warns Dr Kehoe: “My earlier PhD with other virus genomes relied on 24 samples, and that analysis took me about 18 months.”  With supercomputing, Dr Kehoe estimates she could now do that in a week.

The Solution

Pawsey supercomputing was needed to speed up the assembling, mapping and phylogenetic analysis of Dr Kehoe’s genetic samples, which resulted in 13 complete genomes for GLRV 3 and 12 complete genomes for GLRV 1.  Although analysis suggests that all of the known GLRV 1 genomes so far are similar, the number available for comparison worldwide is small.  However, the 13 new GLRV 3 genomes, added to the existing worldwide database of 56, suggest that the diversity of this virus is quite high.

These Australian genome sequences were then used to adjust the testing procedures used for GLRV 3 in WA, changing the primers to ensure that screening for GLRV 3 in Western Australia is up-to-date and less likely to miss local strains.


The DPIRD Diagnostics and Laboratory Service (DDLS) now has improved diagnostic capability for GLRV 1 and 3, and can test large numbers of samples with relatively short turnaround times.   Vineyard owners can now identify and remove infected vines before the virus spreads, and can also be confident in planting clean vines when developing new vineyard blocks.  Dr Kehoe has also developed a portable molecular diagnostic test for those specific virus strains, with the potential to be used in-field by growers and agronomists to identify GLRV 1 and 3 within 30 minutes.  “This would have taken years without the help of Pawsey supercomputing, and I can be confident that my new tests cover all the known diversity of GLRV 1 and GLRV 3, which now includes local WA virus genomes.”

“The new virus genomes will soon be added to the GenBank database,” adds Dr Kehoe, “so they’re available to other researchers across Australia and internationally to include when they design new molecular tests for their local conditions and viral strains, as well as for broader research and monitoring of viruses in our crops that impact on our food security.”

The molecular techniques developed and the workflows created at Pawsey to speed up the analysis required is giving DPIRD the ability to routinely test for more viruses with locally-specific diagnostic tools, not just in vineyards, but also in our grains industry, and in our cucurbit (zucchini, cucumber and melon) growing regions.  By making the vast amounts of information unlocked by next-generation genetic sequencing technologies accessible in realistic timeframes, supercomputing is allowing DPIRD to protect agricultural industries across Western Australia from the impacts of viral disease.

The size of that task shouldn’t be underestimated, warns Dr Kehoe: “My earlier PhD with other virus genomes relied on 24 samples, and that analysis took me about 18 months.”  With supercomputing, Dr Kehoe estimates she could now do that in a week.
Dr Monica Kehoe, WA Department of Primary Industries and Regional Development (DPIRD) ,
Project Leader.