Making research more personal with IBM’s science slams

I always enjoy a good personal story. It’s a nice way to get to know people, and retelling such stories is my favorite part of journalism. So I enjoyed moderating the first IBM Science Slam talks at the recent celebration of the 30th anniversary of the IBM Almaden Research Center.

Almaden, one of 12 IBM Research facilities, is the birthplace of technologies like the relational database, nanotechnology (the ability to position individual atoms), the first data mining algorithms, and the world’s smallest disk drive. Talks on the campus, located in the rolling hills of San Jose, have typically underscored the world’s big problems, like cancer and epidemics, and investigated how to address them with tech.

[aditude-amp id="flyingcarpet" targeting='{"env":"staging","page_type":"article","post_id":2029581,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,","session":"B"}']

But I moderated a section dubbed Science Slams, short talks akin to poetry slams or improvisational talks. Their aim is to inspire, educate, and engage the audience of researchers. The speakers included Almaden researchers Kristen Beck, Rudy Wojtecki, Kun Hu, and Meena Nagarajan.

Each had an interesting tale to tell. Beck talked about keeping the world’s food supply safer. Wojtecki shared how his father’s diagnosis with cancer had affected him. Hu related the impact of the SARS virus in China on her own life. And Nagarajan talked about using supercomputing to help researchers keep up with the huge volume of research in the life sciences.

Here’s an edited transcript of their talks. Over time, IBM plans on doing more science slams at its various locations.

VentureBeat: These science slams are a new kind of storytelling about science. Our researchers are drawing a personal connection between their own lives and the work they do in presentations that are five to 10 minutes long. They’re focused on inspiring, educating, and engaging. Our first speaker is IBM researcher Kristen Beck. She’s going to talk about metatranscriptomics and sequencing the food supply chain.

Kristen Beck: What I’ve come to realize through giving this talk is that metatranscriptomics is a challenging word to say and understand, but hopefully you guys will have a better idea, after this talk, about what metatranscriptomics is. Prior to the break, you heard about the consortium for sequencing the food supply chain from my colleagues. I want to tell you a bit more about the details of the research we’re doing.

At the age of four, my nephew was hospitalized with an ulcer in his bladder. He had several surgeries. During his treatment, he contracted MRSA, methicillin-resistant staphylococcus aureus. Not only did the ailment almost kill him, but he’ll be stricken with this harmful bacteria that’s antibiotic-resistant for the rest of his life. This puts him in harm’s way every time he has an open wound or even a simple scrape.

One of the goals for the consortium for sequencing the food supply chain is to understand the anti-microbial-resistant signatures that occur in food ingredients, using metatranscriptomics. Metatranscriptomics is all of the RNA in an environment from microorganisms, such as bacteria or viruses. This can be used as an identifier of a response to environmental stimuli.

Imagine a jigsaw puzzle with thousands of pieces, each with a different shape, color, and pattern. These pieces are the RNA, the signals a bacteria cell are producing. These sequences come together as a genome to create a picture that means something – the Golden Gate Bridge, say. But just as the bridge does not in isolation define San Francisco, one genome can’t in isolation define a microbial community.

[aditude-amp id="medium1" targeting='{"env":"staging","page_type":"article","post_id":2029581,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,","session":"B"}']

Metatranscriptomics has the power to be exponentially more sensitive and more comprehensive than traditional food safety tests. That’s why, here at the Almaden Research Center, we’re harnessing this powerful opportunity.

Let me come back to the puzzle analogy. If you were describing the skyline of San Francisco, one picture wouldn’t do it justice. You’d need different pictures of different buildings, like the Transamerica Pyramid or the Coit Tower or the Bay Bridge. Take the pieces of all those pictures, put them in a box, and shake them up. Now your task is to identify which puzzle each piece came from and the meaning of those pictures. This is challenging enough with a year of lazy Sunday afternoons to work with. But that’s our task at the consortium for sequencing the food supply chain.

For us, the data we rely on can’t be felt or seen. We only rely on biological sequencing data. While this data may come from microscopic organisms, it’s anything but small. We’re dealing with terabytes of raw puzzle pieces and distilling them into thousands of microbes. We’re looking forward to optimizing this opportunity and using it to apply to the global food supply chain.

Previously, in this field of study, the best practices have only been able to determine the potential functions a microbiome may have. This isn’t as complete a picture as we could be achieving. Through metatranscriptomics and the varying levels of RNA and gene expression, we’re able to understand what these microbes are doing.

[aditude-amp id="medium2" targeting='{"env":"staging","page_type":"article","post_id":2029581,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,","session":"B"}']

If you walk away with one thing from this talk, let it be that through metatranscriptomics, we can understand the functions and what these microbes may be doing. Are they doing normal cellular things, like helping us with digestion or keeping dangerous bacteria in check? Or are they producing potentially harmful toxins?

In order to understand this difference better, picture a roomful of people. If you have a list of the names of the people in the room, you may get an idea of what’s happening in the room — for example, all the attendees are political figures with military strategy experience. This level of information is comparable to metagenomics, the most common method we use to understand microbiomes. However, at this level of information, you don’t understand what these people are actually talking about, what the tone of the room is.

Through metatranscriptomics, we get a transcript of the discussion. We can understand if these are peaceful negotiations, about to reach a resolution, or if there’s a grand disruption, and global war is coming. Being able to discern between outcomes is crucial to understanding a potentially harmful situation in the global food supply chain.

I’m very passionate about this research, because I believe that food is the first step to health for all of us. The microbiomes that live on that food can affect our well-being in an enormous way. By being able to understand metatranscriptomics and the microbes that contribute to those communities, we’ll be able to decrease ill-informed or overzealous initiatives that lead to deleterious side effects, like that strain of anti-microbial-resistant bacteria that affects my nephew.

[aditude-amp id="medium3" targeting='{"env":"staging","page_type":"article","post_id":2029581,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,","session":"B"}']

Globally, every individual is affected by the food supply chain. By characterizing the omnipresent microbes that live on our food, we’ll be able to decrease waste and increase food safety for all.

VentureBeat: Next, we have Rudy Wojtecki, IBM senior engineer and a lithography expert. He’s going to talk about patents, invention, and his surprising path to IBM.

Rudy Wojtecki: Growing up, my father gave me a deep appreciation for math, science, and engineering. He taught me that these were the foundations of a society. This is what we use to build buildings and bridges, understand and predict the weather, and create a truly global society. He also used this as a way to open my eyes to the realm of possibility. You can use these fields to change the world. For me, it was a pathway out of a small town in Ohio.

In high school, my father was diagnosed with cancer. I decided to stay in that small town. In between waiting rooms and hospitals that summer, I had a lot of time to read. Most of my time was spent reading popular science and science fiction novels. These highlighted this idea that you could change a society, reform it, with technology.

[aditude-amp id="medium4" targeting='{"env":"staging","page_type":"article","post_id":2029581,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,","session":"B"}']

One book, in particular, had a dramatic impact on me — a book that described a society changed by something called nanotechnology. It described primarily what could be. It talked about materials that, when you break them, could be re-healed, re-formed, and repair themselves. Nano-scale robots could cure diseases. It also talked about the manipulation of individual atoms to build things like circuits. But it was the first chapter in that book that had the most profound impact on me. It described the pioneering, fundamental work of an IBM researcher named Don Eigler.

I’ll never forget those vivid images of a quantum corral, where individual xenon atoms were moved on a surface to form a ring. For the first time, we could visualize a quantum interference pattern, where these atoms could be moved around on a surface to spell out something like “IBM,” with dimensions that could be measured in atoms across. This inspired me, and so during my college years, I took as many math, physics, and science classes as I could. I wanted to work in this field. I knew nanotechnology was something of the future, something that challenged our fundamental understanding of the world around us.

After graduating, I was able to enroll in graduate school, studying and making simple molecular motors out of chemical synthesis. During graduate school, I heard a guest lecturer named Al Nelson. He was also an IBM researcher, and he described a wild process, something he described as “self-assembly.” It sounded more like science fiction than anything else. He could take a material and controllably cause it to contort, twist, and bend in a shape that he could predict and control.

In his particular application — this wasn’t just an academic endeavour — he actually used this to make readable media by controlling the placement of magnetic nanoparticles on a surface. He also spelled out “IBM” in this case, but with dimensions a little bit bigger than individual atoms. This showed me the breadth of the research that was going on here at Almaden. It showed that there was a wide variety of research topics being investigated. It inspired me to want to work at this place. More than a thousand miles away, in a rural town, I was influenced by the research done right here at Almaden.

[aditude-amp id="medium5" targeting='{"env":"staging","page_type":"article","post_id":2029581,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,","session":"B"}']

After graduate school, I was given the opportunity to work here. Of course I jumped on it. While here, I’ve been able to engage in a number of research projects, everywhere from lithography, where we’re making materials that can be used in the computer industry to reach the five-nanometer node, where circuit components at this scale will be measured in atoms across, all the way to designing things like organic MRI contrast agents, where we can make materials that reduce the toxicity of chemicals used to image the body in an MRI. After three years, I have more than 60 patent applications that span everywhere from lithography to the purification of solutions that could be used in the beverage industry as healthy alternatives to conventional materials.

Fourteen years ago, a book chapter foreshadowed my career here at IBM. What better place to work than the birthplace of nanotechnology? The older I get, the more I appreciate working this field — not only as a way to change and improve society, but as a way to keep the memory of my father alive.

VentureBeat: Next we have Kun Hu, an IBM researcher on public health, talking about disease modeling for global pandemics.

Kun Hu: In 2003, the SARS epidemic spread across an entire country, from the south to the north of China, and then globally. As a college student majoring in electronic engineering, I didn’t understand why we were locked down on campus for so many months. Why was SARS so bad that we couldn’t contain it and prevent it from spreading globally? This inspired my curiosity to find out more about infectious disease — why and how public officials conceive and deploy different intervention strategies.

Many years later, through my scientific training, I learned that following the emergence of infectious disease is not trivial. The establishment of a theory around infectious disease is a complex phenomenon. There are lots of interacting factors, such as the environment where the pathogens and the host are situated. An infected population may not only be human beings — it might also be the mosquitoes that transmit Zika. Also, there are many internal and external dynamics around the affected population. The mass isolation I experienced in 2003 was considered the most effective non-pharmaceutical intervention at that moment to protect a large number of susceptible people in a country with high population density, simply because we had little information about vectors, no vaccine, and no cure.

Several years later, I earned my PhD in the U.S. and joined the IBM Almaden Research Center. As a research scientist, I worked on a project called Spatial Temporal Epidemiology Modeler, or STEM, for short. The goal of mathematical epidemiology is to build a model and quantify the dynamic transmission of infectious disease across space and time. We can use STEM to help scientists and policy-makers quickly come up with a mathematical model so we can fight an emerging infectious disease right away.

Predicting infectious disease is somewhat like computing compound interest. Suppose one day a kid shows up at school with measles. We know that right away, another 13 to 16 kids can be infected. With the STEM equation, you know that the disease can spread exponentially until it reaches all the people who can be infected. Using the STEM equation, you can also envision another scenario where we all vaccinate our kids against the measles infection. This is how we, as scientists, use a modeler to quantify and validate different intervention strategies and come up with suggestions for policy-makers as to what they can do before and during an outbreak.

Armed with the necessary software and various data about things like population and transportation, all kinds of information in STEM, we’re inviting global collaborators to work on and solve some of the world’s real problems, from fighting against dengue fever in Singapore and the Philippines to estimating potential transmission rates of influenza in China to evaluating different vaccination strategies for measles in southern London. We’re participating in IBM’s global brand impact programs on the Zika virus in Brazil.

One important project I want to mention here is IBM’s leadership in the fight against Ebola. The recent Ebola epidemic claimed more than 11,000 lives. Our team at Almaden worked with a global team to build models and come up with important suggestions about policy for government agencies from the U.S. to Australia. Our published results indicated that quick access to resources and human behavior changes could help affected countries control the disease. Our work was so instrumental to the success of our clients and the team that we earned the IBM Research Division Award.

Now, as Jeff introduced earlier on, we’re ready to jump into new challenges and solve some puzzles in the food safety area. Using the grocery-shopping data that you contribute every day, along with a big science approach, we’re ready to accelerate the foodborne disease investigation process. This can potentially impact more than 48 million people in the U.S. I hope to update you about my next scientific breakthrough soon.

VentureBeat: And now we have Meena Nagarajan, an IBM research staff member on IBM Watson health innovations. She’s talking about applying machine learning to cancer research.

Meena Nagarajan: In 1817, a French author going by the pen name Stendhal was touring Florence, admiring art. All of a sudden, he experienced heart palpitations, dizziness, and extreme anxiety, as if he were overwhelmed by the abundance of great art around him. An Italian psychiatrist would later call this phenomenon the Stendhal Syndrome.

Dr. Jerry Avorn, a medical practitioner, in a 2013 article opined that researchers today go through a very similar kind of Stendhal Syndrome, overwhelmed by the incessant waves of information coming at them. He noted that it takes dozens of hours per week just to keep up with information. Sometimes even that is not enough. In domains like cancer research, more than 100,000 articles are published every year.

Medical researchers rely on human cognition, which, as powerful as it is, has limitations: in scalability — how much you can read, collect, and remember — and in bias. Medical researchers are trained to think deeply in one or two areas. It’s hard to take a step back and make truly novel connections, repeatedly and consistently.

Unfortunately, the cost of missing information or going down the wrong path is very expensive in health care. Patient care depends on researchers and doctors being able to assimilate what’s already known so they can follow leads that are more likely to succeed than fail.

So what are we doing to help? Here at IBM, we’re building Watson for drug discovery, a cognitive platform that can read millions of biomedical articles, understand the biology of a disease, and comprehend scientific terminology. Watson can read what’s been written to date and make predictions and hypotheses that are simply hidden connections in biology, so that it can take to the right experiments, the right set of patients, and the right treatment strategies.

This is a story about how we used Watson to help researchers in cancer, and also a story about a protein in your body called TP53 — tumor-suppressor protein 53, also considered the guardian of our genome. When our bodies are faced with genomic abnormalities and our cells are stressed — they’re damaged and need repair — TP53 comes into action. It increases the expression of hundreds of proteins in our body, as if to recruit them to go fix these issues. In short, this protein does a lot in our body to keep us from getting cancer and other diseases.

TP53 gets this call to arms from a different set of proteins called kinases. You can imagine why discovering kinases is important if you have malfunctioning TP53. You can target a kinase with a drug and then essentially ask TP53 to do all the things it does.

When we started working with Baylor College of Medicine, they asked us whether Watson could review articles on kinases and predict new ones. At that point in time, human researchers had identified only about 30 kinases over the last 30 years. The rate of discovery was only about one per year, among researchers all over the world.

We put Watson to the task. We gave it articles on kinases, taught it to read the biology of cancer and TP53, and asked it to predict new TP53 kinases. Watson came back with six strong recommendations for as-yet-undiscovered TP53 kinases. Baylor took these into the lab for experiments, and the results were amazing. Two out of the six came back with strong evidence that they were indeed TP53 kinases. We took Watson and helped researchers accelerate the pace of discovery of TP53 kinases from one a year to multiples in a matter of weeks.

Now we’re applying this technology to 70 use cases in biology. Imagine the pace of acceleration in your domain — materials sciences, clean water, fuel, energy, all domains that are characterized by the same problems of too much information, and that have the same pressing needs for new discoveries.

Perhaps the best compliment Watson has received to date was from one of our partners at Glaxo Smith-Kline. They said, “Watson doesn’t just answer questions. It compels you to think widely.” It’s as if we’re saying, “Look over here, think about this.” That’s the most exciting part of this technology.

When I started at IBM as an intern, close to a decade ago, I was teaching machines to understand slang and emoji. Today, I’m teaching them to understand the language of biology and cancer. When I come to work every day, I’m excited by the common theme that underlies the work we do. When we augment human cognition with machine intelligence, the pace of discovery can accelerate dramatically. That’s our technology moon shot.