Cancer cells within a tumor are not all the same. This variety leads to a large variation in the efficacy of therapies. Enabled by AI, WittGen is looking at the genetics of individual cancer cells to recommend optimal treatment. Founders Yun Rhie, Minwoo Jung, and Minjun Kim gave us a glimpse at their tech and their business plans.
Highlights:
Sal Daher Introduces Yun Rhie, Minwoo Jung, and Minjun Kim
"... The problem you have is that you have to figure out what kind of a tumor you have in order to treat it, to have the right medication to treat it...."
Looking Ahead at What's to Come for WittGen
"... Using sequencing of the RNA, which tells you what genes are being expressed and being triggered, and then you can look at that and then compare that to samples of the cells, looked at in the microscope and so on..."
What Sets WittGen Apart
How the Team Came to Be
ANGEL INVEST BOSTON IS SPONSORED BY:
Transcript of “WittGen Bio”
Guests: Yun Rhie, Minwoo Jung, and Minjun Kim
Sal Daher: I'm really proud to say that the Angel Invest Boston podcast is sponsored by Purdue University entrepreneurship and Peter Fasse, patent attorney at Fish & Richardson. Purdue is exceptional in its support of its faculty, faculty for its top five engineering school in helping them get their technology from the lab, out to the market, out to industry, out to the clinic.
Peter Fasse is also a great support to entrepreneurs. He is a patent attorney specializing in microfluidics and has been tremendously helpful to some of the startups which I'm involved, including a startup that came out of Purdue, Savron technologies. I'm proud to have these two sponsors for my podcast.
Sal Daher Introduces Yun Rhie, Minwoo Jung, and Minjun Kim
Welcome to Angel Invest Boston, conversations with Boston's most interesting founders and angels. Today we are privileged to have with us the team from WittGen Biotechnologies. WittGen is as if you're starting to spell Wittgenstein, the philosopher, it's W-I-T-T-G-E-N and then no stein. The W and the G are capitalized for branding. We're thinking about branding already.
Anyway, the founders are Yun Rhie, Minwoo Jung, and Minjun Kim. Let's do a little bit of bio later on but right now let's talk about the problem you guys are trying to solve. Because my listeners are very eager to hear what companies are doing. Bios can come later. Yun, CEO, please tell us what problem you're solving.
Yun Rhie: Hi, Mr. Daher, and thanks for the invitation. It's such an honor.
Sal Daher: Glad to have you.
"... The problem you have is that you have to figure out what kind of a tumor you have in order to treat it, to have the right medication to treat it...."
Yun Rhie: WittGen is a software leading oncology to doing less, does more. It's a clinical support tool that delivers a precision oncology using AI and mining RNA sequencing data, single-cell RNA sequencing data. The problem we are solving is that currently, the clinical success rate in oncology is very low. Because with current tumor profiling approaches, understanding tumor clone dynamics is very hard. Basically, precision oncology requires molecular profiling.
Sal Daher: Tumor clone dynamics, you mean the reproduction of tumor cells? Are they identical or are they differentiating? Are they evolving within the tumor or is it completely homogeneous? That's what you mean. The problem you have is that you have to figure out what kind of a tumor you have in order to treat it, to have the right medication to treat it. There is heterogeneity, there's differences within the tumor, the types of cells that are multiplying, and figuring that out is a problem.
You believe that with single-cell RNA sequencing, looking at the RNA of single cells that are captured, you can categorize types of cells that you're seeing in the tumor. Then say, these types of cells are amenable to this type of treatment. Is that right?
Yun Rhie: Correct. You point out exactly.
Sal Daher: Just for those who may not be familiar, RNA is part of the genetic code that's used in transmitting instructions to cells and so forth. The DNA stores information, the RNA helps convey that information to make it effective in the cells. Where's the big breakthrough? Is it on the AI side, is it on the wet-lab side?
Yun Rhie: Is at AI side. We developed a cloud-based and AI-driven clinical decision support software for complimentary pre-surgical diagnostics.
Sal Daher: What's the workflow? The oncologist would be taking a whole bunch of samples of the cells and then sequencing the RNA in those cells and then providing that as an input to your algorithm.
Yun Rhie: Yes? They all know this. There's only single-cell RNA sequencing data on the tumor biopsy.
Sal Daher: Of the tumor biopsy.
Yun Rhie: Yes.
Sal Daher: How many different cells would be involved in that?
Yun Rhie: Currently, we curated more than 200,000 single cell samples from 150 patients of the breast cancer.
Sal Daher: In an individual biopsy, how many of the target cells are you going to be looking at, rough numbers?
Yun Rhie: 1,000 single cells.
Sal Daher: 1,000 single cells out of the billions of cells taken or maybe millions of cells taken in the biopsy.
Yun Rhie: Yes, but it varies depending on the sample.
Sal Daher: It's in the order of that. The idea is that the oncologists would get a sampling of all the different cells, sequence RNA of the cells, upload that data into your cloud. Then you would do your crunching and say, this profile of cells may be you're amenable to this type of therapy or that type of therapy. That's the idea, right? As you get more and more cases in your cloud, you're going to have more and more knowledge about this.
That is basically going to be, we hope, your moat to keep competitors out because that's going to be proprietary. You're going to own the data that's coming in. The theory here is that your ability to make good suggestions will be improving over time. How do you get to the point where you're going to have oncologists giving you data if you don't have a training set large enough yet? Or do you have a training set?
Yun Rhie: Initially you got to approach to the researchers at the clinics or some cancer institutions. In the exchange of their data, of their single-cell data, we curate and pre-process instead of them, which means downstream analysis. We take care of their downstream analysis in a brief time with the affordable price. Then by using that strategy, we hope to get enough data to augment our initial machine learning algorithm we have created now.
Sal Daher: When you say downstream analysis, what is it that you mean? Downstream from what?
Yun Rhie: From the sequencing.
Minwoo Jung: Downstream analysis means it is usually using the research-based analysis. When a researcher first get the sequencing data, the raw sequencing data, they need to make that into the count matrix table of each gene and each cell and then sort out some non-significant cells and genes and quality control. Also, they have to do all normalizing, scaling, and cluster each cell types and adaptations. All that kind of thing takes pretty a lot of time for a researcher to do that.
Sal Daher: How is it done now?
Minwoo Jung: For now researchers use, there is some kind of libraries and packages for doing that and code. Usually, they usually use R and sometimes Python to handle their data. Actually, as a data is a very specific data and very large, it takes a lot of time to deal with it, especially when they have multiple patients.
Sal Daher: I'm sorry, you're mentioning Python and R. Python is a programming language. R is a programming language for statistics. What you're doing is you're removing labor from the researcher saying, instead of you hiring a data scientist or someone like that to do that, give it to us. We'll do the number crunching for you, we can maybe even subsidize that with a little-- We have data scientists and all that, we can do that for you. What we get from you is a training set. We will develop and we'll build a training set.
Minwoo Jung: That's right.
Sal Daher: Then how do you close the loop to know that if you came back with the data from the downstream processing and then the hypothesis as to which therapy should be applied. How do you close the loop on that? How do you know what therapy was chosen? Is that part of your deal with the researcher, that they would suggest this therapy? Minjun Kim, who's on the biology side of things on the life science side, McGill University doctoral candidate, please continue.
Minjun Kim: The drug recommendations are made usually for the clinicians because they need to have better decision-making for the patients. Those informations are for the clinicians. For the researchers, they need to process the data and make the things more organized so that they can dig into their field of interest.
Sal Daher: The clinicians are the ones taking samples?
Minjun Kim: Yes.
Sal Daher: These presumably clinicians are involved in research. The researchers would have the samples of the clinicians, your involvement here is that, "Hey, we're going to make your life easier with the downstream processing. After you get the stuff sequenced, we'll do the downstream processing for you and let us have the data."
Then presumably when the clinicians decide on a course of treatment, you're going to close the loop by finding out the results and how that treatment went? Is that part of the deal that you have with the researchers?
Minjun Kim: Like the proposed should be a bit different. Even though the researchers could be clinicians this way, the proposal of research is not always consistent with the clinical aspect. Sometimes they dig into the biology, underline the phenotype picker aspects or something similar.
Sal Daher: I'm sorry. They use the phenotype to make decisions about drugging, about what drugs to use?
Minjun Kim: It's a bit different. Researchers, not always they are focusing on the real clinical aspects. Even though they are MDPhD, their research is not always to treat the patients. Sometimes we need to understand more about the cancer--
Sal Daher: It's not necessarily closing the loop on clinical results, but just figuring out what kind of cell you're dealing with. When you see the genetics and then you see what's actually expressed in the sample, you know what's in the sequence and the phenotype is what's actually there and you can observe it in some other ways. Then you can see what genes have been expressed and which haven't and so forth. You will learn things about the behavior of the cancer short of an ultimate clinical outcome.
Therefore, by having richer information about the population of cancer cells and the tumor, you know the RNA, sequencing of it, but then the stuff doesn't look the way you expect it to look. Presumably, you're looking at pathology and things like that and you're saying, "Ok, wait a second. It doesn't jive, there's something more at play." So you have richer information.
What indications do you have that there is a commercially viable business here that you guys can make money from doing this? Where is it that you think people would pay you or when would they pay you to do it? Maybe they won't pay you now but when you have two years of training data, what's the thought in terms of a business model here?
Looking Ahead at What's to Come for WittGen
Yun Rhie: We'll provide much more reliable and comprehensive information of cancer intra heterogeneity to clinicians, minimizing trials and errors for treatment optimization and making cancer care more cost-effective. What makes WittGen distinctive to the customers are our solution can track cancer origin, whether it would be primary or metastasis. Or where it came from in the organ of the human bodies.
Also, currently only histological way, that means post-surgery, exists in the molecular subtype inference, but we do it at the cytopathology phase. That makes huge difference we believe. Also, we can integration of single nuclear type variants, SMB, and gene copy number variants for genetic mutation identification as well. Probably, we believe, and we'll make the clinician can find first-in-class, best-in-class, therapy recommendation ultimately.
Sal Daher: At the moment, what you're doing is you're offering convenience to researchers in exchange for the training set. Is that fair to characterize it as that?
Yun Rhie: The reason why we approach the researchers first is that there is two main goals. One is to get the data much more than now we have, and the other one is that validate our solutions from the researchers, and that we'll lead to the clinicians' visibility.
Sal Daher: All right. Just to make sure that everybody understands, you're hoping to get the training set. What's the other thing that you're trying to get from your collaboration with researchers again, if you can restate that?
Yun Rhie: For the validation purpose. Once we get the validation from them, the clinicians get-- We get the attention from the clinicians.
Sal Daher: Now, the validation, what form would you expect that to come in? People say, you're a SaaS platform and you guys provided information to me that was really valuable, I want to give you more data sets. At that point, you're going to say, "it's validated, I'm going to start charging you". Basically, if people are very excited about what you're doing, then at that point you got to start saying, pay. Because you guys have to support yourselves, have to make money, and so forth. Is that what you're saying?
Yun Rhie: Yes. It's to generate the demand from the clinicians. Resulted more than 250,000 single-cell data for the machine learning inference. We already trained more than 80,000 single-cell data for molecular subtype and great prediction from the breast cancer patient.
Sal Daher: Would you gimme those numbers again, please?
Yun Rhie: 80,000.
Sal Daher: Data from 80,000 different biopsies?
Yun Rhie: Yes.
Sal Daher: That's a big step.
Minwoo Jung: I think we have to like clarify here.
Sal Daher: Yes. Minwoo Jung.
Minwoo Jung: 80,000, different cells.
Sal Daher: 80,000 different cells. How many different patients is that?
Minwoo Jung: 80,000 different cells from about 60 patients.
Sal Daher: 60 patients. Then we get the order of magnitude, a thousand-something cells per patient on average.
Minwoo Jung: That's right.
Sal Daher: That's good to get things a little bit more concrete. This is interesting because I'm an investor in a company called Picton Health. At Picton, I think it's public, it has a very large data set for training on images of skin lesions that have been diagnosed. The founder, shout out to Susan Conover and Pranav Kuber, co-founders, they've been able to convince clinicians to turn over a very, very large number. They have many different clinicians, a large number of diagnosed skin conditions for them to use for their training set.
I don't see that they're offering any benefit to the clinicians at all, other than just being good human beings. [chuckles] What you guys are doing is that you're actually, you're making their lives easier. You get to be a good human beings, researchers, and plus, we're going to give you more time to play with your kids or something. Or do something else other than just be number-crunching all the time.
Which is not what you guys do anyway, you're cancer researchers, you want to do other stuff. I get the sense that you guys have something to offer in terms of doing this crunching of numbers for existing researchers. Very good. You guys are very, very early on. I find this really compelling. There's some people in my group who do AI, at Walnut Ventures, who look at AI and we'll see if they'll take a look at this. I'll send a link to the interview to a couple of the AI guys there and see what their thoughts are.
Yun Rhie: At this moment, we actually achieved a decent amount of the, a decent number of the accuracy that like more than 90% accuracy of subtype prediction of breast cancer and cancer grade prediction from top 20 gene expressions at the cellular level. The grade is used to help predict clinical outcome and figure out what it looks like.
Sal Daher: The data sets that you're matching is, on one side you have the RNA sequencing data that shows you the gene expression, the genes that are being triggered. Then you have histology data that shows what the phenotype of the cells present are within the tumors. Your predictions are about 90% accurate. Your model is able to predict that match 90-something percent of the time. It means you guys are cooking with gas, you've got gas in your stove. You're not just cooking with no gas.
Yun Rhie: You can say that again.
Sal Daher: All right. [chuckles] You guys, as I say, we're pretty early on. You still have to- - You're going to incorporate and set up. What are you looking for right now? What kind of help would be most valuable to you at this point?
Yun Rhie: Most importantly you're seeking an angel investment or a proceed investment.
Sal Daher: Angel investment, that means you have to be incorporated. You have to--
Yun Rhie: Oh, sorry, I need to clarify. The incorporation is in progress actually. We're working with that.
Sal Daher: Oh, it's in progress.
Yun Rhie: Yes. In progress.
"... Using sequencing of the RNA, which tells you what genes are being expressed and being triggered, and then you can look at that and then compare that to samples of the cells, looked at in the microscope and so on..."
Sal Daher: You're working on that. Eventually, you're going to be raising money. What is it that you would be doing with the money if you are able to raise that?
Yun Rhie: Oh, with the money we're going to hire the key researchers in machine learning art to accelerate the progress of the algorithm development and evolution. That's the main point.
Sal Daher: Very good. All right, we've talked a little bit about what WittGen Biotechnologies is working on, namely, using artificial intelligence, machine learning, to better categorize the variety of cancer cells that exist within a tumor. Because surprise, surprise, not all tumor cells are alike in a tumor. Some tumor cells let their hair down, some are clean cut, some do this, some do that, there's that heterogeneity within tumors.
Using sequencing of the RNA, which tells you what genes are being expressed and being triggered, and then you can look at that and then compare that to samples of the cells, looked at in the microscope and so on. Between those two, you can triangulate and figure out what's going on, what kind of cells are present, and what the RNA signature means about those cells.
You're hoping that if you do that enough, you're going to be able to just look at the RNA, the sequencing data, and say, you got this type of cancer, use this type of anti-cancer agent. It's going to be 30% better, 100% better, 10 times better than current standards. That's the hope. Yun Rhie is nodding yes. Let's do this. Is there anything else that you want to add to the state of the company and what you're doing, what you're working on at this point? Because the next half, I want to get a little bit more of how you guys came together, biographical, that kind of thing.
Yun Rhie: It sounds good. Currently, within a month our web application portal for the clinicians, which runs our machine learning algorithm will be released supported by Amazon AWS investment. They granted us 26k in cash.
Sal Daher: Congratulations.
Yun Rhie: We just kicked off the meeting and to raise to build. It takes just one month. Our clinicians or researchers can see our product as soon as possible.
Sal Daher: My question was also, where is it that you have found the researchers to collaborate so far?
Yun Rhie: We got four different institutions. First is Amsterdam University Medical Center, we are working with oncologynomics, with principal investigators.
Sal Daher: Yes, Albert Einstein.
Yun Rhie: Then there's two Korean institutions, which are Yonsei Medical University and Kaist. Also, we had a discussion with a potential collaborator, in terms of using our tool, with GSK and GlaxoSmithKline cancer genomics r&d center.
Sal Daher: You have three research centers and then a strategy you're interfacing with, GSK, GlaxoSmithKline, a big pharmaceutical company?
Yun Rhie: Yes.
Sal Daher: Very good.
Minwoo Jung: Can I add?
Sal Daher: Minwoo Jung, please.
Minwoo Jung: I just want to talk about the technical expansion of this technique. We are capturing millions of cells, it would be billions of cells from a lot of patients. Capturing the RNA expression and usually, their target is decided by RNA expressions. When we capture all kind of RNA expression from the single cells of the patients, we can figure out new drug target to inhibit the cancers like severity or development. That is why we can get involved with pharma companies.
Sal Daher: Okay. What you're saying is if you're able to match the RNA expression to particular drug targets, which would be manifested in the types of cells that are present, you can match the drugs more closely? This has drawn interest from strategic players.
Minwoo Jung: Yes. That's right.
Sal Daher: Feel free to correct me.
Minwoo Jung: Yes, overall is correct but to be a little more specific, if we see the gene expression of the different cancer patients' data, there are "gene markers" to differentiate these patients. Some patients have those genes, for example, some patients have those genes, and these people who had these genes had very high severity, and some people who didn't have these genes had low severity. Maybe when we inhibit the expression of these genes, then they would get okay.
Sal Daher: Okay. What you're hoping to do is you're hoping to suppress the expression of genes that are associated with very severe manifestations of the cancer.
Minwoo Jung: That is a part of drug development. Because we have all those gene expression data, we can work with pharma companies to new drug development or the enhance of existing drugs.
Sal Daher: Just for those who may be a little bit lost to this gene expression, a gene may be present but the gene is not always on. DNA has tons and tons of information, and most of the genes, the genetic information that is carried into the DNA is not necessarily activated. When we need it, certain types of genes are called into being. Very mundane things for example. Here's the mind-blowing example of [chuckles] genetic expression.
If you are exercising, you run, doing your everyday run-of-the-mill exercising, there's a set of genes that's involved with maintaining your muscles, rebuilding your muscles, and so forth. They've done tests with the legs of frogs and they have exercised the frogs way beyond the normal amounts of exercise that these frog legs are used to doing. Then the recovery phase for these frogs was they triggered a different set of genes from the genes that are normally present in the recovery of the muscles. The idea here is the pushing beyond your normal boundary, kind of like, what's the expression, getting out of your comfort zone?
When you're within your comfort zone, there's one set of genes that maintain your muscles. When you get outside the comfort zone, there's a different set of genes that are triggered to rebuild your muscles. Then you get additional muscle, thicker muscle cells, you get more muscle cells, and all that stuff. This is part of weight training, is triggering the expression of genes that build the muscle instead of just sustain the muscle. The body has tons, and tons of genes that are not acting, so thus gene expression. I just thought it would be an important side here for people who may not be familiar with this. Very good.
Minjun Kim: I can add something
Sal Daher: Minjun.
Minjun Kim: Yes. The point here is a tumor is not a single clone organism, it's a multiclonal mass of diverse lineage. We believe the main reason that the tumor is now responding completely to the treatment-- Cancer treatment initially shows the effect so they do shrink, but at the end, they are very likely to develop additional tumors or grow again, something like this. We believe the main reason for this is the clonality of the phase.
That's why we need to do single-cell RNA sequencing for the tumor diagnosis. Because before this technology is available, it's been several years, but before that we used the conventional operative, like [inaudible 00:29:07]. Multiclonal mass of diverse lineage in one cell and profile it and then show it. Single-cell technology is doing the same thing for it in the tumor and then we can get the information for each cell. So that we can decide what treatments should be made to get rid of the whole tumor.
Sal Daher: To adjust for the heterogeneity, the different types of cells inside the tumor, you're saying it's not monoclonal, it's not all identical. So then you can optimize for the mix of cells inside the tumor. Yun, wanted to say something here.
Yun Rhie: Yes. I just want to add onto that Minjun is saying. Just want to wrap up our--
Sal Daher: Sure, please. Please, Yun.
What Sets WittGen Apart
Yun Rhie: What makes us different from other players in the market is that we make a software that can provide clinicians with high-resolution cancer profiles at the decision-making time before surgery. I think that makes a huge difference. The clinicians currently heavily rely on histopathology they use. We use solid tumor biopsy samples beyond psychological analysis and build optimal neoadjuvant treatment strategy.
Sal Daher: What you're saying is they're taking samples and then they're fixing these samples and looking at them through microscopes and trying to look at the shapes that they have and so forth to categorize the kinds of cells that are in there. That's very limiting in the sense that you can't do that at scale, it requires a lot of time on the microscope looking at a lot of cells.
Whereas if you're doing the single-cell approach that you're talking about, you're able to get almost a statistical approach to it. You have so many of this type of cells, so many of that type of cell, and so forth. It becomes a more scalable process, am I right to say that, instead of the traditional pathology method of taking the sample and fixing it and then looking at it and so on?
Interestingly, there are developments that are happening right now in AI at looking at the histological samples, of these samples have been fixed. Path AI is the big name in that, they've just made a lot of strides in that. The whole universe is looking very closely at what they're doing. This is a different perspective saying, "Instead of just looking at that, let's also look at the sequencing data and enrich your pathology study with the sequencing data." It's a way of tying the two together.
Yun Rhie: When it comes to precision medicine or precision oncology, these days digital pathology image analysis and next-gen sequencing data analysis are primary drivers. Then next-gen sequencing is divided into two parts again, DNA panel sequencing, and RNA sequencing. Most players these days such as [unintelligible 00:32:22] belong to the digital pathology image analysis and DNA panel sequencing market.
The reason why we more focused on the single set RNA sequencing unlike them, because the various image analysis and DNA sequencing data analysis has a X-ray effect and CT effect respectfully. We believe the single sequencing RNA data analysis has some MRI effect in diagnostics. That's what I'd like to stress out.
Sal Daher: Okay. You're using MRI and X-ray as metaphors, in a sense that--
Yun Rhie: Yes, an analogy.
Sal Daher: Yes. An analogy. X-ray is static, the MRI, you can see motion, you see more.
Yun Rhie: Detect more.
Sal Daher: Detect more. Very good. Let's do a little brief promo for the podcast, and then the second part we'll speak briefly about how you guys came about, how the team came about. Now, if you're enjoying this conversation, it's really we are in the edge of science here doing stuff that's really new and that's really groundbreaking. If you enjoy this kind of stuff, you can promote that by first subscribing to the podcast so you can find it again.
You'd be surprised how many people stumble on the podcast and then never find it again because there are a gazillion podcasts, lots of people sending stuff. Subscribing to the podcast means it shows up in your feed every week when the podcast gets launched and then you get to choose which particular episode you want to listen to. Do subscribe to the Angel Invest Boston Podcast.
The other thing that would really help us, if you're really psyched about a particular episode, you can upvote that episode by leaving a rating and a written review. It doesn't have to be very long, it doesn't have to be a lot of writing, a little bit of writing, a little pithy, "Love the work you guys are doing on single-cell RNA sequencing and tying that to heterogeneity of cells in cancer. I love it because I'm in that field or whatever. I love that podcast." Make that comment. The subject line is everything in reviews and then the review itself, less important. Make that effort.
What that does is it helps us get found on Apple Podcasts or Google Play or whatever platform you listen to. Please, if you're inspired by a conversation that we have, and I know that our people, I know the startups have been started because of conversations that have happened here. Take your time, a little time and leave this a review and a rating. By the way, it doesn't show up for 24 hours, for example, in the Apple Podcast platform. I think there's some kind of duration that they do. It takes time to show up. Don't expect immediate gratification, but eventually, you'll be rewarded with seeing your review and ratings up. Yun Rhie, Minwoo Jung, and Minjun Kim, how did you guys come together?
How the Team Came to Be
Yun Rhie: I think it's been a year and a half passed. We first met in a organization is called K-BioX. It's a science and biology scientists community for Koreans in the world.
Sal Daher: Physically, Minjun is in-
Minjun Kim: Montreal.
Sal Daher: - Montreal. Minwoo, where are you sitting?
Minwoo Jung: I'm an Indiana, in Purdue University.
Sal Daher: Hey, a Purdue connection. This podcast is sponsored by Purdue University so good to have the connection. Yun, you're in Boston?
Yun Rhie: I live in Berkeley, California.
Sal Daher: Oh, that's right. I should also take this moment to thank Marco Salvalaggio, my colleague at Walnut for connecting us. Because Marco connected with you guys because his connection at Berkeley. Marco is one of these migrant angels. When the weather is nice in Boston, he hangs out here. When the weather turns nasty, he goes to California and hangs out around Berkeley, the phenomenon 21st Century migrant angel. You've hooked up with a Marco Salvalaggio, my Walnut colleague at Berkeley. That makes perfect sense.
Yun Rhie: Yes. We are in the Sky Deck incubating program. We met Marco through the office hour for his advice. Thanks, Marco.
Sal Daher: Great guy. It doesn't show but he has a lot of gray hair. He has a lot of experience starting businesses and so forth. Very good. Connected via this set up for connecting Korean ex-pats. Very good. How did the idea of starting the company come about?
Yun Rhie: It's a long story. I used to be an investment banker for about 12 years and then I moved to Berkeley to pursue my MBA at UC Berkeley HaaS school of business. After graduation, I tried to several, a few startups in FinTech and semiconductor area using AI, but I failed twice. I tried to find out the opportunity doing some leveraging AI technology. I found the great market in biotech after all and also because I also concentrate, I studied with the concentration on biotech commercialization.
Sal Daher: Biotech what?
Yun Rhie: Commercialization. I started to study and learn more about the bio side.
Sal Daher: What is it that gave you this hint that this could be fertile soil to till?
Yun Rhie: There a lot of hurdles and challenges in oncogenesis. While learning the biotech area aspect, the knowledge about the biotech, I just wanted to try and validate my ideas how could-- What if I combined AI and biotech, what would be the result? That's where I started. Then I started to get some advice from people like Minjun and Minwoo, K-BioX, and other PhDs.
I translated into the asking phase to the writing a journal phase with them. Asking them to, "Let's validate my idea and combine your thoughts and then make it some paper to publish. Then let us validate the opportunities." Then at the phase three in the early of this year, and I asked Minjun and Minwoo do this business together.
Sal Daher: Okay. It started out as a research project, put together a paper, and then eventually said, "Hey, it looks like we can actually start a company." Let's map it out. Yun Rhie from the investment banking background goes to get an MBA at the Haas business school at Berkeley, Minwoo Jung, you're in artificial intelligence, and Minjun Kim, you're in the life sciences. You're getting your PhD in exactly what major?
Minjun Kim: My major is human genetics and I'm focusing on the cancer genomics with the bioinformatics operations.
Sal Daher: Okay. It's kind of computationally intensive side of genetics.
Minjun Kim: Yes. I'm doing both, wet lab and dry lab.
Sal Daher: Okay. The usual life science company that I see is like a technical founder and then trying to do everything. This is a multidisciplinary startup because you're doing the AI and you're doing the wet lab side of things, and you have someone with some experience in the investment world.
Okay, guys, if there's anything else that you want to add at this point, let's put it in, and then we'll wrap up the interview.
Yun Rhie: One thing I missed to tell you is that our IP, we are start to file the provisional IP for our technologies.
Sal Daher: That's very good. That's very good.
[music]
Sal Daher: Well, Yun Rhie, Minwoo Jung, and Minjun Kim, I thank you three for making time to be on the Angel Invest Boston podcast.
Minjun Kim: Thank you for having us.
Sal Daher: Excellent, Minjun.
Minwoo Jung: Thank you.
Yun Rhie: Thank you. It's such an honor to meet you here and a chance to promote WittGen Biotechs.
[music]
Sal Daher: The three co-founders of WittGen Biotechnologies. This is Angel Invest Boston, I'm Sal Daher.
[music]
Sal Daher: I'm glad you were able to join us. Our engineer is Raul Rosa. Our theme was composed by John McKusick. Our graphic design is by Katharine Woodman Maynard. Our host is coached by Grace Daher.