Satvik Dasariraju ’23 Published in Bioengineering Journal

A scientific research paper by Satvik Dasariraju ’23, Detection and Classification of Immature Leukocytes for Diagnosis of Acute Myeloid Leukemia Using Random Forest Algorithm, has been published in the most recent issue of Bioengineering. The Third Former took a few moments to discuss his work.

Why did you decided to pursue this opportunity via via iResearch (iRI)?
I was motivated to apply because of my passion for learning new things, especially in science and mathematics. I applied to a few summer programs and ultimately selected iRI because of the well-designed science research curriculum, as well as the program director's experience and tradition for excellence. I also credit and am very grateful to Dr. [Daniel] Concepcion, my IBES (Second Form Biology) teacher last year, for furthering my interest in biological sciences and writing a letter of recommendation.

What was the application process?
The required materials for the application to iRI were fairly standard for competitive summer programs. I submitted five essays, my transcript, a letter of recommendation, and a list of accomplishments and extracurricular activities. Students are evaluated by the program director and teachers based on their demonstrated passion for science, collaborative ability, writing skills, and of course grades and awards.

How were you selected for this particular project?
I was matched to the research field of computational biology, specifically biomedical image analysis, based on the interests I outlined in my essays and on my past experience in STEM.

I wanted to pursue a study in computational biology after taking a Harvard course on edX and learning how to employ the powerful capabilities of computer science to solve medical problems. I was assigned to biomedical image analysis because iRI believed it would be a good fit based on my strengths. Later on during the program, I identified leukemia as a disease where biomedical image analysis could come in handy for early detection and classification into subtypes.

Would you briefly describe the work you completed?
My project, which I completed with the assistance and guidance of a student researcher at Stanford and the iRI director. While I actually led the project and did most of the work, their guidance was crucial.  

The work consisted of four main components. First, I created a segmentation algorithm for blood cells, which basically isolated the nucleus and cytoplasm of over 1000 blood cells from a publicly available data source. Second, I extracted cell features from the isolated nucleus and cytoplasm (eg. nucleus to cytoplasm area ratio, nucleus perimeter, etc). Third, I trained a type of machine learning algorithm (called the Random Forest) to detect immature blood cells (which are only found in diseased conditions such as leukemia) and classify the blood cells into four types. My model was able to achieve over 90% accuracy for both detection and classification. Finally, I performed mathematical calculations to determine the most important cell features for the task, which can be used by researchers in future studies as a baseline. Ultimately, my model can be used as a support tool for doctors to efficiently and accurately diagnose leukemia. All the work was done virtually and the project was coded in Python.

Is the project now complete?
With the publication of the paper, the project is now complete. However, scientific research is never complete and I hope researchers can use the cell features I calculated to be most important in future studies. I hope to pursue similar projects in the field of biomedical image analysis in the near future.

What do you feel you learned from this experience? How do you feel it will help you in your Lawrenceville studies?
Other than the vast amount of information I learned by reading countless scientific papers while I planned and set up my project, the most important lesson I've learned is that nothing can limit someone that is passionate and focused on the topic. I accomplished a lot more this summer than I could have expected, especially as a high school student.

I hope that what I've learned about biology and computer science with benefit me in advanced STEM courses such as Honors Biology, particularly the research-oriented classes because my research experience should be helpful.

How does it feel to be a published researcher?
This is the first "real" research paper I've written and got published. I feel that writing mini-essays last spring in IBES helped me understand scientific writing and its differences from writing in humanities. The publication process consisted of an editorial check, two rounds of external peer review by three independent experts, a decision by an academic editor, and a decision by the journal editor. The journal asked for numerous major and minor edits based on the comments of the peer reviewers, which I found to be insightful for the improvement of my paper. 

After I submitted the paper, I was definitely nervous and on edge for the next two months, but it was extremely rewarding when I got the paper published, especially since I'm the lead/first author. I got over the initial excitement in a few days because I know I still have a long way to go.

For additional information, please contact Lisa M. Gillard Hanson, director of Public Relations, at lgillard@lawrenceville.org.