Spotlight: Josh
Josh Espinoza interviewed by Sam McCabe
Josh has an incredible breadth and depth of knowledge that is immediately recognized by his intelligent messages on our Discord. Always willing to help answer any questions new members might have, or add his perspectives on elaborate discussions from NewAtlantis regulars, the community would not be the same without Josh. It was a pleasure to learn about just how multifaceted and interesting his life has been, and I hope one day I can have a gorilla bestie like him (read his Spotlight to learn about his gorilla bestie, Denny).
Where are you currently living and where are you from?
I was born and raised in San Diego, CA. I studied abroad in New Zealand and lived in Santa Cruz, CA for a few years but found my way back to where I started.
You have quite an impressive educational background, when did you find your calling to pursue metagenomics and machine learning?
My research trajectory has been more of a curious wander than a straight path; taking opportunities as they arise here and there. When I was an undergrad at San Diego State University (SDSU), my original major was Astronomy but I changed it to Cellular Biology right before the first semester started. My mom was really supportive of whatever I wanted to pursue but she definitely hinted towards me going into pre-med; biology was something in between. It seemed like studying the origins of life would be more rewarding than the origins of light at the time anyways. Definitely the right move looking back 15 years later.
My first research opportunity was translating recordings of bird songs into discrete digital units (a fancy way of saying that I was dragging a mouse over a spectrograph and saving files while listening to music on Pandora — a dated term). Despite what you might think, sitting for hours in a windowless lab cataloging bird calls gives you a lot of time to daydream. I always thought about how epic it would be to travel to remote locations and actually collect the samples myself or how I could go about interpreting what the birds were communicating with each other.
Around this time is when I decided to study abroad in New Zealand where I took an ecology course that involved a lot of fieldwork in some incredible nature reserves. I’ve always been into Nature (e.g., David Attenborough documentaries and exploring tidepools) but this experience was when I started to get perspective on the complexity of an ecosystem. It was like a fractal of information where you would look into a question and once you zoomed in realized it was full of minutiae that warranted their own branching questions.
When I got back to the US, I took a Bioinformatics course and an Ecological Metagenomics course in the same semester. My professors were awesome and kudos to them for addressing my (probably ridiculous) curiosities during their office hours. Ultimately, this experience led me down the rabbit hole of computational biology and how it could be applied to understanding the influence microbes have on ecosystem (in)stability. At the end of the semester, I wanted to go deeper so I started a joint senior research project with my professors from the previous semester. This allowed me to continue the development of my programming skills while trying to figure out what microbes were actually doing in the ocean.
While I was in my senior year at SDSU, I enrolled in a Stem Cell and Regenerative Biology course which sparked my curiosity again. Again, I had a lot of questions about concepts tangentially related to the topics we would discuss in class so office hours were where I could elaborate. During office hours, my professor strongly encouraged me to apply to the stem cell internship program through the California Institute of Regenerative Medicine (CIRM) as he was on the board. I was a bit reluctant because I applied the year before and was denied but applied anyway. I got accepted towards the end of the semester but there was a caveat to the program, you had to be an undergrad to be in it and I had enough credits to graduate. Though it was paid and I didn’t really have anything planned for the aftermath of SDSU so I delayed graduation to see where it would lead. In this experience, I realized that culturing cells and being not so kind to mice (to say the least) wasn’t for me but fortunately I had some incredible advisors in the lab that gave me opportunities to use my nascent bioinformatics skills in their research which is where I was able to find my place.
From there I got accepted into a Bioinformatics Master’s program at the University of California, Santa Cruz (UCSC). The campus is nestled in the middle of a coastal redwood forest (with caves on campus) with remote coastal cliffs and weird dive bars within a 10-minute drive from campus. At the time, I was really into the idea of minimal synthetic cells so I joined a synthetic biology-focused lab that also dabbled in metagenomics. I took a Systems Biology course where our main project was to enter into a Kaggle competition developing machine learning models to classify cancer phenotypes; this is when I realized that predictive modeling and machine learning was my thing.
In between years at UCSC, I got accepted into a Microbial Ecology internship at the J. Craig Venter Institute (JCVI) and learned to explore the landscape of computational biology in the context of microbial ecology (e.g., ocean microbiomes and human microbiomes primarily). After graduating from UCSC, I completed a joint Biotechnology Ph.D. program through JCVI and the Durban University of Technology, South Africa (DUT) where I got to dive deeper into biomarker discovery, predictive modeling, and ecological network analysis; I’ve been at JCVI ever since. At JCVI, I have been extremely fortunate to have an incredible mentor with interdisciplinary expertise in machine learning, marine ecology, and human biology; he has been instrumental to my development as a scientist.
Long story short, I’m a very curious person and I’ve tried it all but metagenomics and machine learning were, in a sense, always calling to me.
For all those that aren’t quite as familiar with your areas of expertise, how would you describe them in simpler terms?
I guess I’m a jack-of-all-trades but in a nutshell, I would describe the predictive modeling part of machine learning as a way to train a computer to make predictions based on patterns in the data. Explainable artificial intelligence is the approach I use for my development as this places emphasis on a predictive model’s performance (e.g., accuracy) as well as its human interpretability. Feature selection (e.g., biomarker discovery) goes hand-in-hand with predictive modeling as giving the models better options to choose from will give you better predictions in the end.
To better understand feature selection, imagine you’re looking for the tangerine in your backpack and start fumbling around with your hand. You feel some soft fabric; that’s not it, probably your beanie. You find some flat smooth surface; that’s not it, that’s probably your sketchbook. You find some round object but it’s fuzzy; that’s probably your tennis ball from physical therapy. You finally found some smooth and textured round object; that’s probably your tangerine. Identifying the informative features like shape and texture that help with figuring out whether or not you found the tangerine is the general idea of feature selection.
This concept of a feature being informative for a prediction is important for finding biomarkers diagnostic of a disease or microbes that are indicative of a changing environment. Doing this helps ensure your machine learning model will not be overwhelmed by irrelevant or unimportant information and will be able to make more accurate predictions. Further, this helps with developing models that generalize to other datasets and not overfitting to the input data — a term called overfitting. The algorithms I develop identify/rank those features and investigate how those features interact in a way that can provide insight into one system relative to another such as a healthy versus diseased patient or a warming ocean versus a stable ocean.
Is there any specific research or topics you’ve studied that you are really passionate about and think other people should know about them?
Throughout my scientific journey, I’ve had some amazing collaborations with researchers from different cultures studying a wide range of topics amounting to 22 peer-reviewed research papers and 1 NASA whitepaper. There are 3 topics I’m passionate about that come to mind and have given my research a greater sense of purpose.
For the first, I was punted a project studying plastic-degrading microbes in the ocean that was in a sort of stasis due to the lack of metagenomics workflows that could reliably handle eukaryotic organisms like algae and diatoms. I love diving and exploring the coast so plastic has been a huge concern for me ever since I could remember. I started working on my VEBA software suite as a side project during the beginning of the pandemic so I could finally figure out what was going on with the plastic colonizers in this dataset. The software suite I developed turned into something much more comprehensive and I was able to apply the methodology to every metagenomics dataset I’ve had since its inception.
The second project would have to be an antimicrobial discovery pipeline we and our collaborators developed to determine the mechanism of action of crude extracts with unknown antibiotic activity isolated from producer strain microbes in soils; a major bottleneck in drug discovery pipelines. With this project, I was able to develop my expertise in explainable artificial intelligence through hierarchical predictive modeling and feature selection algorithms for biomarker discovery. Our group also discovered darobactin, an antibiotic with novel activity produced by a bacterium living in the gut of a soil-burrowing parasitic nematode.
The last project would be a project where we and our collaborators identified and characterized beneficial microbes that could help corals survive in warming conditions. This was challenging since it involved eukaryotic gene modeling and coral microbiome analysis for a coral whose genome wasn’t available.
Weird microbes are everywhere doing weird things that could be useful for humans.
What is the challenge you would most like to solve?
The challenge I would like to solve in this lifetime would be to transition social, corporate, and political interests from short-term exploitation of natural resources towards long-term sustainability paradigms that value Nature. This isn’t something one person can do but our collective influence may shift the scale towards the latter and, by extension, the trajectory of humankind and the fate of the only planet we are certain has developed/can sustain life as we know it.
You consider yourself an environmentalist. What does that mean to you? And how would you advocate for more people to become environmentalists?
My notion of being an environmentalist has evolved over the years. When I was growing up, it was about appreciating any Nature I had access to from flipping over logs in my backyard to the creek of runoff at the park down the street to the tide pools during a new moon’s low tide. As I grew up and become more independent (i.e., I could actually travel places), my idea of what it means to be an environmentalist started to focus more on experiencing exotic areas; though, as I’ve gotten more responsibilities over the years (e.g., married, a career, and 2 cats Keven & Eleven), I’ve learned to revert back to my childhood mentality of finding awe in both the simple and the exotic (when I can). Appreciating each nuance from microscopic insects on a leaf to swimming with leopard sharks while dodging stingrays to 2 am meteor showers on a desert cliff brings you into the moment. Those moments of appreciation and awe when in natural systems bring you closer to Nature as you realize that you’re an extension of the environment and the environment is an extension of yourself. Our actions and intentions have ripple effects so our values aggregate outwards towards the rest of society, corporate interest, and government policy.
So what does being an environmentalist mean to me? It means not only appreciating Nature but having a sense of responsibility to protect these natural systems; systems that not only give us awe but also sustain our livelihoods and are largely responsible for our own existence.
My core philosophy is that there’s only one habitable planet we have access to and humans are (arguably) the only known instance of advanced intelligence (i.e., influence over the Geosphere > Biosphere > Technosphere, refer to this paper for insight into the concept of planetary intelligence). I believe that it’s our responsibility as humans to preserve the natural ecosystems that have provided us with the opportunity to develop intelligence and, in doing so, we can simultaneously progress our unique instance of intelligence sustainably with Nature.
How do you think that your area of expertise is going to affect the world in the next 10 years?
The tech industry is developing incredible machine learning and analytical frameworks at an increasing rate. The cost of DNA sequencing is rapidly declining and sequencing technologies are becoming more efficient. Appreciation of “The Microbiome” and its diffusion from microbiology to ecology to common knowledge is promising as this means more public interest and more funding for impactful research.
The synergy between machine learning and metagenomics, the study of microbiomes from a genomics perspective, is nascent but rapidly progressing (e.g., CheckM2, VAMB) and is more firmly rooted in mathematical theory (e.g., compositional data analysis). In the next 10 years, I can imagine this duality maturing and becoming as coveted as cancer and longevity research by both prospective scientists and investors. If the explainable artificial intelligence methodologies developed in the tech and biotech industries can be leveraged towards addressing climate change and biodiversity collapse in collaboration with domain experts, then we may actually stand a chance for a sustainable future that is mutually beneficial for both the environment and humanity. I am hopeful that my research can play a role (even indirectly) in progressing the world toward this type of future.
How did you become aware of New Atlantis? And how is this community different from others you’ve been a part of?
As I mentioned earlier, I feel like environmental metagenomics has always been calling to me and Jayson Gutierrez (@jagut) reaching out after my VEBA software suite was published is a testament to this. We had a few conversations over the course of a few weekends and the mission of New Atlantis seemed to align with the direction I’d like my research to go towards in the long term; that is, leveraging cutting-edge machine learning technologies for investigating topics in climate change, biodiversity, bioprospecting, and plastic degradation.
The New Atlantis community is different from other communities I’ve been a part of because we all have the same goal of preserving the environment and biodiversity while open-sourcing our expertise in ways that are beneficial to the overall mission. I’ve never been a part of a community with such a wide range of expertise from scientists to journalists to design experts. Also, Discord is weird and new to me but I’m into it.
Do you have any memorable moments in the ocean?
The first one that comes to mind was spearfishing with my best friend in Catalina Island, CA where we found an octopus and a moray eel in the same session (not for hunting, just admiring). We noticed that the eel had a 3 pronged hook sealing the top part of its jaw with the bottom part; it was completely exhausted waiting for its own slow demise. We worked together taking turns as one of us would hold the eel, and the other would try to unhook its jaws. Eventually, we got the hook out and watched it swim away to live another day. Another time on the Big Island of Hawaii a couple of my closest friends and I were diving around and swam towards a pod of at least 30 spinner dolphins. We would dive down deep and they swam with and around us without any hesitation, just playing and blowing bubbles around us like we were part of their clique. It was surreal and felt like a dream. Another collection of times would be in New Zealand exploring the southerly ocean with completely foreign fauna both at sea and on land, but to be honest, every time I float around in the ocean and connect with the moment is a memorable experience that fills my cup. Also, not sure if it counts but I participated in a FishHackathon competition at the Monterey Bay Aquarium and while I was there between coding sessions I saw a pod of orca hunting a gray whale calf in the bay…sad but epic and definitely memorable nonetheless.
What is your favorite marine animal?
Too many creatures to pick just one…I’d have to split this into invertebrates, vertebrates, and plants. Invertebrates it would be tough but I guess I’d have to go with jellyfish as there is just something really mystical about them especially when you find them in the wild. For vertebrates, without a doubt, it is going to be marine iguanas. Not sure if it was my childhood obsession with Godzilla or the National Geographic Reptiles and Amphibians VHS I would watch on repeat but marine iguanas have always been numero uno. Lastly for plants, it’s going to be kelp for sure. Diving through kelp forest when we get good visibility in Southern California is one of my favorite ways to spend a sunny day and definitely one of my favorite ways to explore Nature.
Lastly, what do you do when you’re not doing research?
When I’m not programming or read/write/exe papers, I’m usually either rock climbing, in/at the ocean, meeting up with people at a brewery or cafe, or looking for an epic lookout to vibe for a bit. My wife and I try to make our way out to Joshua Tree as often as we can and are regulars at the San Diego Zoo; Denny the juvenile lowland gorilla and I are low-key besties at this point. If I’m feeling particularly inspired then I like to draw stuff like landscapes, maps, natural history, and chimeras from time to time. I used to produce a lot of music (@O.rka on Soundcloud) but it’s slowed down a bit as my research has progressed (especially during my Ph.D.). Mostly, I like having a good time with chill people in cool places.
Learn more about Josh through his Github, Google Scholar, Instagram, or LinkedIn.
NewAtlantis seeks to address the twin challenges of climate change and biodiversity loss by aligning community, government, industry, and individual benefit with the improving ecological health of our oceans. Subscribe to our newsletter on our website, or join our Discord to learn more.
