Karthik Jagadeesh is a CS Ph.D. student in the Bejerano Lab. He has a B.S. in EECS and B.A. in Applied Mathematics from UC Berkeley. He is building statistical models to identify disease causing mutations and computational tools to improve rare disease diagnosis.
Let’s start by having you to tell us a bit about yourself, personally and professionally.
I am a Computer Science PhD student in Gill Bejerano’s research group. My research is focused on studying the human genome to build computational tools and statistical methods to better interpret patient’s genomes and diagnose Mendelian diseases. Prior to joining Stanford, I did not have much exposure to computational biology, but I fell in love with genomics after joining the Bejerano Lab and seeing how we can extract meaning from genomic code. I completed my undergraduate degree from UC Berkeley in EECS and Applied Math, where I had brief research stints in databases and computer security and completed an honors thesis exploring approximation algorithms in Applied Mathematics.
I grew up 20 minutes south of Stanford in Saratoga, CA and am a Bay Area native. I enjoy playing tennis, hiking, and traveling in my free time.
Why did you become a scientist? Did you want to be a scientist as a child?
I spent my childhood at the center of the internet technology revolution in Silicon Valley. Growing up in this environment, I was impressed by the impact people here had on the local community and the world. I soon realized that fundamental research in science and technology is what allows this impact to be possible. Starting in undergrad, I explored various research topics and was particularly excited by data and applied machine learning problems. After coming to Stanford, I found that computational tools have the ability to transform our understanding of the human genome. From the compilation of my undergrad research and work experiences, I have refined my interest in computational genomics and specifically human disease genomics.
Can you tell us about your current research and what you hope to achieve with it? You could start by listing 3 words you think best represent or embody your research.
Mendelian disease diagnosis.
My research goal is to bridge the gap between genotype and phenotype, and understand the ways in which genetic variation contributes to phenotypic diversity and ultimately human disease. Mendelian diseases are diseases that are caused by mutations to a single gene. As typical individuals have around 4 million variants genome wide, we aim to develop computational techniques to find that “needle in the haystack”. Using data from next-generation sequencing and tools from computational biology and machine learning, I have worked on projects ranging from machine learning for variant pathogenicity prediction to cryptographic techniques for genome privacy.
M-CAP, a machine learning classifier, reduces the variants of uncertain significance in patient genomes by 60% while maintaining high sensitivity, and has become widely incorporated into exome analysis projects and pipelines. The scores are available through the popular ANNOVAR tool as well as through dbNSFP. Additionally, thousands of users have downloaded our VCF file of exome wide scores and hundreds of users query our online web search tool every day.
Another recently published project on genome privacy introduces a secure framework to compute flexible functions over genomes from multiple individuals while preserving patient’s genomic privacy. Genome privacy is becoming more important, especially as we start to understand more about the genome and how it encodes function. It is critical to build secure methods for genomic analysis as this allows individuals to continue to share their genomes while maintaining anonymity.
Recent Press on genome privacy:
- Wired – https://www.wired.com/story/to-protect-genetic-privacy-encrypt-your-dna/
- Scientific American – https://www.scientificamerican.com/article/cryptographers-and-geneticists-unite-to-analyze-genomes-they-can-rsquo-t-see/
Were there people in particular to whom you would attribute your professional success?
There are many people that have helped me get to where I am and for whom I am deeply grateful, including my family, my lab mates and especially my mentor Aaron Wenger and my advisor Gill Bejerano.
Aaron taught me a lot about genomics, specifically in the context of disease, and he helped me frame my first project, and structure a concrete approach to tackle it. I have continued to employ a lot of these strategies I learned from him on projects I worked on later. Aaron and Harendra Guturu (another one of my mentors) both built many of the computational tools and data processing pipelines that are a foundation of the medical genetics work I have done.
Gill gave me an opportunity to join his lab though I came with no background in biology. One of my favorite parts about working with him is the interesting discussions and brainstorming sessions that we have. He constantly pushes me to think deeply about my research and asks questions, forcing me to have a thorough understanding of the underlying methods. Gill also has a detail-oriented approach to research, writing, and presenting which has taught me not only how to approach research but to also package and present my work so it can be understood by multiple audiences.
Research is a truly collaborative process and working with the members of the Bejerano lab and seeing how ideas and tools have been passed down and improved generation by generation has really influenced my experience as a PhD student.
What are your future plans? Where do you see yourself professionally in the next 5 or 10 years?
I am planning to graduate in the summer and am currently applying to postdocs. I hope to continue improving our understanding of the genetic basis of diseases, specifically through expanding interpretability of the noncoding genome and complex diseases. Eventually, I hope to become a PI with a lab focused on developing computational methods to improve the understanding of the genetic architectures of human disease.
Can you speak a bit to the role you see CEHG playing on Stanford campus?
CEHG is awesome! CEHG organizes many interesting talks and lectures throughout the year, such as the CEHG symposium, and these events provide a great opportunity to learn about new topics as well as meet other people working on related topics.
What advice would you offer to other grad students or postdocs who are considering pursuing a similar educational and career path as you?
In general, I feel there are a lot of interesting opportunities at the intersection of fields. As a young scientist, it is good to spend time developing a strong foundation and skills in two diverse areas. This allows you to have the expertise to build novel solutions to problems by connecting ideas from different areas – something not many can do.