Data Science – ACENET

A Prize-Winning User

Erin Johnson wouldn’t have won the prestigious Steacie Prize, awarded to early-career scientists, in 2021, without the Digital Research Alliance of Canada’s high-performance computing and the services of ACENET.

“That’s a completely fair statement,” says Johnson, who was an early adopter of high-performance computing, dating back to her student days at Carleton University, Queen’s University and Dalhousie University, as well as her post-doctoral work at Duke University. She’s now a professor and Herzberg-Becke chair in theoretical chemistry at Dalhousie University.

Asked how she would explain her job to a guest at a cocktail party, Johnson says she studies intermolecular interactions within the materials all around us.

“These are the weaker interactions between molecules as opposed to the stronger bonds within a molecule,” Johnson explains. “So, we would, for example, at a cocktail party look at the interactions in beer or wine in a glass — those types of interactions within a liquid or a gas or within a molecular crystal.”

She does this to predict such properties as reactivity, hardness, and conductivity, among others.

“Any chemical observable is something that we can predict through this type of modelling,” she says.

Johnson’s lab, however, might surprise those who envision her working in a chemistry lab full of beakers and bunsen burners. Rather, it’s a computer lab from which she’s developed methods for modeling that are some of the most accurate and efficient available. It was those methods that won her the Steacie Prize.

“We then take those methods and apply them to problems in chemistry,” she explains. “One of the particular problems we’re focusing on is the problem of molecular crystal structure prediction or how molecules would come together to form a 3D solid.”

She likened the ways molecules come together to the ways in which Lego bricks do.

“There are many ways you could conceive of molecules coming and packing together, but not all of those are going to be stable,” she says. “So the challenge is trying to predict how they will actually pack in a solid.”

Once you solve that puzzle, there are applications across several industries, one of the most recognizable being pharmaceuticals.

“If you are producing a drug and you want it in pill form, you want a solid but soluble state,” Johnson says. “When you try to formulate new drugs, you want to screen for all the possible polymorphs (transformations to another form) and find out which ones are the most stable. What you don’t want is for it to easily form a particular polymorph that’s not a very stable one and then change over time so that it’s no longer soluble.”

These methods also have applications for electronics, but instead of looking at solubility, the property of interest would be conductivity.

“Like the solubility, the ability of charge to flow through a material is also dependent on the particular polymorph” Johnson explains. “You can think of many applications where the different solid state properties between polymorphs would affect whether a material was promising or not.”

Johnson says her work would be “completely impossible” without ACENET’s services.

Making Data Meaningful

Whether it be for a community project to determine which water is safe for drinking, or a hospital-based project to determine the most efficient and effective ways to treat mental illness, Trishla Shah is an expert in taking the data we collect so easily today, and making them meaningful.

As the IT research scientist at the Nova Scotia Community College and a PhD candidate in computer science at Dalhousie University, Shah focuses on designing solutions by analyzing massive amounts of data. For example, Nova Scotia’s water comes from various places, including freshwater lakes, but also possibly from contaminated lakes, she says.

“We want to educate the organizations taking care of our water resources so we can see how we can preserve the well water and how we can make sure the water from the fresh lakes is not contaminated,” Shah says, adding that many of her projects come to her after receiving funding from a national social innovation fund.

Private IT companies that want to design and test a product from scratch also often approach the research scientist.

“We are being approached by startups, by midsize companies and by large-scale companies,” Shah says. “We are being approached by organizations that take care of Indigenous communities. One project we’re working on involves artificial intelligence, Indigenous communities and Halifax’s storied Bluenose ship.”

For another project, Shah is helping small- and medium-sized enterprises identify the right clean technology investments for their business. Her lab is building a data repository to provide solutions to scale access to low-carbon emissions for businesses in the agri-food, construction and manufacturing industries. The repository is integrated from various sources using web-scraping and text-mining techniques. The parameters of the data repository will be set to allow an algorithm to forecast carbon emission reductions and analyze the return-on-investment of various interventions.

Shah and her team get involved from the beginning, first collecting the data, then building a database, preprocessing, storing and managing the data in a way that ensures they’re secure and making sure there’s no breach. From there, they can make models of the data so they don’t have to analyze them manually.

Shah hasn’t used the services of ACENET directly, but she has circulated its many programs among the information technology students at the college.

“We have a lot of academic programs that have IT in them,” she says. “And students might want extra practice, or students might need to have something like a prerequisite. With ACENET, the courses are really applicable and I like to circulate them among students who need additional practice on that topic. So that’s my interest with ACENET. I’m a fan.”

In addition, Shah is in the process of designing her own student training module that has an online portal where students can enrol themselves in training that is focused on real-time projects.

“I have proposed to the committee that’s approving my student training module and funding that ACENET’s courses be prerequisites.”

Supercomputers at the Frontline of Climate-Resilient Forests

When he was introduced to forest genetics as a 23-year-old researcher in India, Dr. Om Rajora became obsessed and devoted to forest genetics research. “Sometimes, even at the cost of my health,” he says, laughing.

Decades later and half a world away, he is still going at it, and busier than ever. He heads the forest genetics and genomics lab and is a professor of Forest Genetics and Genomics at the University of New Brunswick, where he joined as the Tier 1 Canada Research Chair in Forest and Conservation Genomics and Biotechnology. He also coordinates the Population, Ecological and Conservation Genetics research unit at the International Union of Forest Research Organizations, and he developed and published a pioneering book series called Population Genomics.

Through his work, Rajora seeks to reveal the secrets hidden within the genes of our woodlands, offering unprecedented insights into resilience, adaptability, and the preservation of biodiversity (genetic diversity) in the face of a rapidly changing world. His research has encompassed a wide range of forest genetic and genomics topics and fields, which besides contributing to basic science, can be used for conservation of genetic resources and sustainable forest management.

One of those areas involves analyzing the genes, gene expression, and metabolic pathways of trees under ambient and changing climate conditions. This in turn gives hints to how forests respond to environmental changes. By sequencing and analyzing the genes expressed under different scenarios, Rajora aims to identify the genes, biological processes and molecular functions affected by climate change, including those related to photosynthesis and stress response. “For example,” he explains, “this could provide us information about what traits could be affected with elevated levels of CO₂.”

Rajora’s research obsession and a strong work ethic have their limits, however, and today’s rapid gene sequencing techniques have created the need for advanced methods of data analysis and storage.

In a recent project, Rajora and his post-doctoral fellow, Dr. Rajni Parmar, needed to undertake extensive genetic analysis of the red spruce transcriptome and chloroplast genome, and identification and characterization of genes and pathways expressed differentially in response to climate change conditions. This required significant computational resources. In the past, they relied on personal computers for certain analyses, but the advent of next-generation sequencing—rapid DNA/RNA sequencing techniques—and the accumulation of massive amounts of data necessitated the use of supercomputers or computer clusters. ACENET and the Digital Research Alliance of Canada have been instrumental in providing these capabilities.

Dr. Serguei Vassiliev, a research consultant with ACENET, played a pivotal role in the study. He delved into the project, learning the fundamentals of computational bioinformatics, installing software, and creating parallelization schemes to process large problems simultaneously. He also trained the group in essential computational techniques, including genome assembly and annotation, transcriptome construction, statistical analysis of gene expression, and protein interaction network identification.

The outcomes were remarkable: computation time was slashed from weeks to hours, the group acquired advanced computational skills, and they published an article in the International Journal of Molecular Science, with another in progress. Vassiliev’s significant contribution earned him co-authorship on the publication.

“The knowledge provided by this research is beneficial to the global plant genomics community,” explains Rajora. The variations found in gene expression clue scientists into what genes may allow forests to survive adverse conditions. Through international collaboration, scientists can pool their resources, data, and insights to develop innovative solutions that can be applied globally in the face of climate change.

From Megabytes to Megafauna: Driving Whale Conservation to New Depths with High Performance Computing

“I think whales are kind of like dinosaurs,” says Tim Frasier. “Almost everyone is interested in them at some point in their childhood, but some of us don’t grow out of it.”

Frasier, now a biology professor at Saint Mary’s University in Halifax, Nova Scotia, is following his dream. His lab studies genetic variations in whale populations to fuel conservation efforts using high performance computing (HPC). Their present focus is on the North Atlantic right whale and Saint-Lawrence beluga.

By observing the variability of an individual’s genome over its lifetime, Frasier and his students determine how inbreeding and traumatic events, such as ship strikes or entanglement in fishing gear, make individuals and populations less able to recover and reproduce.

The Frasier Lab sequences whale genomes from minuscule pieces of skin collected humanely by field teams. They then compare the variation in gene expression in healthy and injured whales, the life histories of which they know thanks to data collected by researchers through photo-identification over the last several decades. All these data provide the basis to understand how the cumulative effects of inbreeding and multiple stress factors can cause individuals to not reproduce or live as long.

Frasier specifies, however, that these analyses involve millions of DNA fragments, which take up huge amounts of both storage space and computing power. “You just can’t do it on a normal computer.” Access to ACENET and the Alliance’s supercomputing facilities are instrumental in enabling his lab to tackle complex genetic analyses that would otherwise be impossible.

Frasier gets asked a lot how this kind of information translates into helping conservation. One way is by influencing policy. Existing legislation imposes limits on different industries according to their impact on whales. For example, a North American right whale hit by a ship or entangled in a fishery somewhere along the east coast of the US and Canada could trigger either a ship slowdown in the area or completely shut down that fishery for the season. He explains that these triggers might only be quantified based on whether the whale died or not, “like a yes-no question.” But if Frasier’s research can show that these incidents change their reproductive success for years afterward, then it would also demonstrate that current measures are vastly insufficient to help populations recover and thrive in the long term.

Frasier’s work both advances our understanding of whales and provides a framework for devising more appropriate measures to encourage population recovery. It also highlights the importance of computing resources in modern biology research and conservation. “Without ACENET, we wouldn’t be able to do this work,” he says. “It’s just such a great resource that we have, and many of the geneticists in Canada that I know feel that way.”

Building Better Brain Imaging Analysis Tools

Erin Mazerolle uses brain imaging to understand, for example, what’s happening in the brain of someone with Alzheimer’s disease.

“A lot of the work that I have done is to try to do a better job of understanding and diagnosing neurological diseases, and to take really advanced cutting-edge research imaging techniques — new ways to image the brain — and apply them for the first time in patient populations,” says Mazerolle, assistant professor of psychology and computer science at St. Francis Xavier University.

Over the course of this research, Mazerolle noticed that she and her peers don’t always do a good job at making sure their results are reproducible in other studies.

“If I do a study and publish it, and then another researcher reads it, can they get the same result?” she asks. “This is the part of my work where ACENET really comes in. I have a study going on right now in which we’re trying to figure out how the different ways that we analyze data are impacting the reproducibility of our results.”

She says that when analyzing brain imaging data, there are dozens of steps researchers take, and each one involves deciding which parameter to use at that juncture. And ultimately, all of those decisions impact the final result.

“There’s an analogy that we use called the ‘garden of forking paths.’ I make a decision and go down one garden path and I never know what my results would have looked like over on the other path,” she says.

“So we’re developing software that researchers can use to simultaneously analyze the imaging data using a whole bunch of different paths. That’s a really major computational challenge because these datasets are very large and it takes a lot of computational time and power to get to the end of each of these paths.”

She says the clusters at ACENET reduce the computational challenge to a reasonable time frame, but even with those high-performance computing resources, it can still take a month to do the analysis for a single study.

“It would be completely unfeasible for me to do it without ACENET,” Mazerolle says.

“We hope that by looking at all of the outcomes of the different paths, we get a better sense of whether our results are real, or if the ones that we like only occur down one or two paths,” she says. “Or, do the 200 paths actually look totally different and we never would have known that if we hadn’t tried this multiverse analysis?”

For her multiverse research, she’s using data from the Alzheimer’s Disease Neuroimaging Initiative, an international data-sharing database that looks at Alzheimer’s Disease, cognitive impairment and healthy normal aging. In particular, she’s comparing super-agers (who are in their 80s or 90s and have the memories of 20- and 30-year-olds) and comparing them to people who are aging normally. This is the study that will pilot the tool she hopes will help further brain imaging research across Canada.

Timing Travel with Precision

Davod Hosseini is a good guy to know if you’re a commuter. That’s because his “operations research” or OR, tries to solve complex problems in transportation planning and logistical management.

“The core process of OR involves identifying an issue, formulating a mathematical model to represent it, gathering and analyzing relevant data to understand its behaviour, and, finally, providing actionable recommendations by finding optimal or near-optimal solutions,” explains Hosseini, assistant professor of management science at Saint Mary’s University’s Sobey School of Business.

Specifically, Hosseini leverages OR to optimize routing, scheduling and vehicle assignment in two types of transportation systems — regular systems, in which he looks at transportation costs and customer satisfaction, and in riskier transportation systems involving transportation of hazardous materials (hazmat) via railroads, in which he aims to minimize risk.

“The optimization process in both cases is performed by considering factors such as traffic patterns, distance, service and delivery times, demand forecast and vehicle capacity, while ensuring all operational constraints, government regulations and customer requests are fulfilled,” Hosseini says. “However, the inherent variability in these factors adds a layer of complexity to the model, making it more challenging to achieve an optimal solution.”

“Companies such as Amazon or FedEx leverage advanced route optimization to plan daily deliveries for thousands of packages,” Hosseini explains. “The system considers factors such as delivery windows, traffic conditions, package size and driver location to create efficient routes and meet tight delivery deadlines.”

Another application is home health-care, where providers visit multiple patients daily. The system can account for patient location, appointment times, last-minute appointment cancellations, new patient requests, care urgency and unexpected traffic delays, to name a few, and then generate route planning that minimizes travel time, ensures on-time visits and maximizes patient experience.

On the hazmat side, one could imagine a use for chemical manufacturers who want to transport their products by rail. The software considers such factors as population density along the route and prioritizes less populated routes in case of an accident, weather conditions and track maintenance.

Hosseini says he uses ACENET services because he usually has large datasets, numerous constraints and variables in the mathematical model, and complex optimization algorithms.

“Traditional computing can become slow and impractical when dealing with vast amounts of data, such as railroad networks with thousands of tracks and intersections or logistics scenarios involving millions of delivery points. HPC provides the processing power to handle these large datasets efficiently.”

He says he has benefited from ACENET staff’s expertise in supercomputing to migrate his workflows from his university desktop to an HPC cluster.

Without ACENET’s tools, Hosseini says his work would take weeks or months to complete.

In Search of Apple Perfection

Zoë Migicovsky has the secret sauce to help farmers get more fruitful crops. The assistant professor of biology at Acadia University and Canada Research Chair (Tier II) in Agri-Food and Sustainable Agriculture works her magic by combining genetic data from fruit crops and agriculturally important trait information about the plants and then analyzing the data so plant breeders can make predictions about what their plants will produce in terms of fruit.

“Agriculturally important traits would be things such as when an apple might ripen or what its aroma might be,” Migicovsky says. “With an apple seedling, you’re going to be waiting four to seven years for there to be enough fruit for you to have a meaningful evaluation of what that fruit is like.”

There’s an enormous resource investment in that seedling in terms of breeding the plants, fertilizing, watering, managing for pests and disease and pruning, to name a few. “Then at the end, most of the plants won’t have desirable traits,” Migicovsky says. “That will be true regardless, but if we can make predictions about some of those traits early on, we can reduce the number of plants that need to be culled at a later date and narrow down which are more likely to be desirable plants later on.

She sums up the problem she’s addressing by quoting from an October 23, 2023 Financial Times article in which a breeder started with 90,000 trees. Of those, “Only 357 varieties made it to a second round of trials, 18 went to a third and 13 reached the final round.” The article states that apples are like diamonds in that way.

In order to do her work, Migicovsky needs computational resources that will handle large genomic trait datasets and link them together. “The plant breeder isn’t going to screen for 200,000 genetic markers. They would like to only screen for a couple, so we need to know which ones are the best for them to do that. If you have a quarter of a million columns to compute, you’re not going to do that on your laptop,” Migicovsky says. “We need computational resources like those available through ACENET and the Digital Research Alliance of Canada.”

The more genomic data she has, the more likely she is to find good predictors, but the more genetic data she has, the more computational resources she also needs.

Using ACENET resources saves her money in her research budget and enables continuity in her work. It’s a shared system, so her students are able to access her lab’s files and software. “It’s helpful because students are only there for a relatively short time so this allows for one student to pick up where another left off.”

Cracking Cannabis Codes

David Joly studies the interaction between plants and micro-organisms.

“We are looking at what genes make plants more resistant to disease and what genes make them more susceptible,” explains Joly, a biology professor at Université de Moncton. “On the pathogen side, we’re trying to determine what genes make a pathogen aggressive with a particular plant and what makes the pathogen detectable by that plant. Plants have an immune system and are able to recognize certain molecules from pathogens, triggering a defence response, a little like we humans do.”

Joly works mostly on cannabis and says some plants are more resistant than others — again, being comparable to humans. Some humans, for example, seem to get the flu every winter, and some simply never seem to get sick.

“If we focus on plants that are more resistant and compare them to plants that are susceptible, can we see differences in the genes?” he says, using tomatoes as an example. “We could take those more resistant plants and use them in a breeding program, crossing them with plants we know produce really juicy tomatoes.”

Joly says he’s starting from the beginning in many ways with cannabis because Canadian researchers have only recently been allowed to study it.

“I have to stick to what’s been authorized to work with,” he says. “So we have to gather as many different plants as possible and test them in our growth cabinets, and then we can sequence their DNA and ultimately use ACENET resources to look at their differences.”

He says he and his team need to screen millions of “letters” in the cannabis genome, and on a small computer, that would take weeks.

“So that’s where we use resources that are available from ACENET,” he says. “What I like about ACENET is the training they offer to take students from zero knowledge of bioinformatics to slowly making them more comfortable with bioinformatics coding. You need to be able to program and code and ACENET teaches them that. I can help them, but we often take advantage of the training from ACENET.”

He says ACENET has been especially instrumental with his undergraduate students, who are almost always new to bioinformatics.

“Even at the graduate level, I have students who arrive here and don’t know much about bioinformatics, so ACENET’s training is still useful there,” he says. “Then we access the different tools and software so we can analyze the data. Every time we encounter problems, the ACENET people are always very useful in helping us find the problem. Sometimes it’s just a semicolon in the coding and they’re patient enough to help us find it.”

Joly says his work would be “very difficult” to do without the services of ACENET.

“You can wash the dishes manually, or you can use the dishwasher, but if you use the dishwasher, you can wash way more dishes in a given amount of time and get other things done while that’s happening,” he says.

Making Generative AI Sound More Human

Uyen “Rachel” Lai is looking into differences between the way AI algorithms that generate text, such as ChatGPT, express themselves compared to humans. The goal is to improve the usefulness of generative AI models in producing content.

“For example, my preliminary work suggests that generative AI models produce text that is more lexically diverse, though possibly less focused,” Lai says. “I want to make generative AI models that produce more “human sounding,” natural or realistic text.”

Lai, a fourth-year computer science honours student is working under the supervision of Paul Sheridan, assistant professor at the School of Mathematical and Computational Sciences at the University of Prince Edward Island. Even as an undergraduate, Lai has travelled as far as Tokyo to present a paper on her work with Sheridan Stable — Sheridan’s professional wrestling-inspired name for his “plucky” group of undergraduate researchers — which focuses on natural language processing and text analysis. The team is trying to understand how the two worlds of classic statistics and deep learning are connected.

“Deep learning inherits a lot of ideas from classical statistical methods. We’re looking at those connections and then mapping them out, which we hope will lead to insights about how to use all the statistics knowledge the world has accumulated over the years to do deep learning in more interpretable and efficient ways.”

Lai is setting the standard in that team of researchers.

“Rachel is an exceptional student who’s on a good roll right now,” Sheridan says, “and there are lots more great students on the way who are hungry to make names for themselves.”

Lai uses ACENET resources to generate text using AI models, and she has also trained some of her fellow students on how to use them.

“We need a lot of text to compare lexical diversity in human-written and AI-generated text,” Lai says. “The compute clusters come in handy, helping us generate a massive amount of text. By using these resources, we can generate what we need in a few hours at most. Without them, it would take months or years to do our work.”

Sheridan says that as the program grows, the team will be generating vast volumes of text that will require hundreds of gigabytes which would be impractical to store on a regular computer.

“We need to use GPUs to generate text,” he says. “GPUs are in high-demand for deep learning applications.”

Lai lauds ACENET’s support services, which she says are also very responsive.

“If one person can’t help, they will find us someone else who has more expertise in the problem we are trying to solve,” she says.

Predicting the Future of Fisheries

Ian Bradbury uses DNA technology to understand what aquatic species — whether Atlantic salmon, cod, crab or lobster to name just a few — Eastern Canada has and how they might respond to stressors such as climate change.

“We look at how things are adapted to the environment using genetic and genomic tools and then we use machine learning and climate models to look at how they might respond in the future,” explains Bradbury, a research scientist with the Department of Fisheries and Oceans and an adjunct professor at Dalhousie and Memorial Universities. (Genetics is the study of how genes work while genomics is the study and mapping of genomes, or the full set of genetic instructions for an organism.)

The goal is to be able to make projections in terms of the rate of change happening with a specific species, as well as how it will respond to climate change and how that might impact fisheries and other stakeholders using those resources in the future.

The classic example, he says, is Arctic char in Labrador, which is culturally important for Indigenous groups along the coast, ecologically important because it’s the dominant freshwater coastal fish species there and notable because it’s at the southern portion of its range in Labrador.

“So it would not be surprising that climate change might be pushing it northward,” Bradbury says. “And we’ve done a lot of work over the last couple of years, some of it using ACENET, to understand how Arctic char in Labrador and north of that, are adapted to their climate and then how climate change might affect that.”

Bradbury is building a baseline or map of genetic variation in Arctic char. His projections suggest that the Arctic char’s range will start moving north, meaning the southern portion of Labrador will no longer be suitable for char, which will have implications for the people living in the area and for the ecosystems that remain there.

Bradbury and his team need ACENET because genetic tools generate massive datasets — multiple terabytes in fact — and since data management and data analysis are most of what they do, they couldn’t do their work without ACENET.

“We don’t have access to the computational power that my students would need to do these sorts of analysis,” Bradbury says.

The students he supervises are sequencing the entire genomes of aquatic species and then analyzing them for differences among individuals and populations.

“We’re making associations with climate on a set, and we’re doing projections for future impacts,” Bradbury says.

While there are other options out there, he says, ACENET is particularly useful because it’s accessible to students. His research team includes students at all levels and postdocs at both Atlantic campuses.