Interview with Upendra Kumar Devisetty

0
40

Upendra Kumar Devisetty is the author of Deep Learning for Genomics; we got the chance to sit down and find out more about his experience of writing with Packt.

Q: What are your specialist tech areas?

Upendra: Bioinformatics, Genomics, Machine Learning, Deep Learning, Distributed Computing.

Q: How did you become an author for Packt? Tell us about your journey. What was your motivation for writing this book?

Upendra: My journey with Packt started when I was asked to review their book “Machine Learning in Biotechnology and Life Sciences”. When I saw that invitation, I had no hesitation in accepting it because of my interest and background in Machine learning in the Biotechnology space. I was inspired by what the author has done in the book and reviewing the whole book made me wonder how awesome it would be if I can do the same. So, after review, I made up my mind to write a book in an area that I was good at. In addition to taking inspiration from the author of the book who I reviewed by also partly comes from my interest in teaching. I taught lots of workshops on Bioinformatics, Big Data, and Docker, and developed a course on Big Data Fundamentals in Data Camp which was taken by more than >10K students. I am a firm believer that I can understand concepts better if I can explain them in a way that the readers find useful. This thought made me explore the opportunities to become an author with Packt. Initially, I expressed my interest to author a book combining PySpark, a distributed computing framework, and Genomics but after a lot of brainstorming of ideas, we decided on Deep Learning for Genomics. In the end, it was a fitting topic and title because of my passion for Deep learning and background in genomics which worked out well. My motivation for writing this book mainly comes from the fact that there is no proper source of knowledge on Deep Learning in Genomics. Deep learning for genomics applications is already revolutionizing many domains such as healthcare, agriculture, clinical, pharmaceutical, and so on. Currently, a lot of resources are available for understanding deep learning but very few resources exist for a proper understanding of deep learning applications for genomics to solve real-world problems. The intersection of Deep learning and genomics is little explored through books because of how technical and complex the concepts are, and it can be quite intimidating for beginners to enter this field. Through the book, my goal is to provide a single source of information that every genomic researcher Data Scientist or any research scientist can bank on to understand the basic concepts of Deep Learning and how it can be leveraged for genomics applications in life sciences and Biotechnology industries. The book is meant to provide a one-stop solution for applying deep learning techniques to address some of the most challenging problems in genomics.

Q: What kind of research did you do, and how long did you spend researching before beginning the book?

Upendra: Honestly, it didn’t take much time for me to start the book because I was very clear from the beginning about what I wanted to cover in the book. Being a Bioinformatics Data Scientist, I have a good background in Deep Learning and coupled with my strength training in genomics, I didn’t have to spend a lot of time coming up with an outline for the book and getting started with the process. I was told by the Packt team that my outline for the book was well structured with proper coverage of all relevant topics in the very first draft which was enough to validate that I was going in the right direction. Of course, I did little research to find out if there are any books but unfortunately, there were none and so it made it easy for me to use my thought process for the book. As soon as the outline was approved, I started writing the chapters without spending much time researching. Overall, less than 2 weeks were spent on researching before beginning the book. Most of the time was spent on writing the book, reviewing it, and creating real-world studies.

Q: Did you face any challenges during the writing process? How did you overcome them?

Upendra: I would be lying if I say the journey was smooth. During the whole process, I faced so many hurdles. Juggling between a full-time job and writing a book was extremely challenging especially toward the end when the turnaround time is very short. Most of what I wanted to cover in the book is very technical and so I had to spend a lot of time breaking down the concepts and making it easy for potential readers. The field of Deep learning is itself very complicated and coupled with the fact that I was trying to integrate that with genomics made things very hard. In summary, I had to break down the concepts not just for one field but for two in this case. The other challenge is coming up with genomics case studies as close to real-world problems and leveraging deep learning to address them was also a herculean task. I overcome all those challenges by sticking to a routine for chapter writing. I dedicated 2-3 hours every workday and 8-12 hours during weekends and of course, holidays are a bonus for me. This allowed me to make incremental progress every day without worrying about spending the whole day writing. Working every day for 2-3 hours may not seem a lot, but you will be amazed by how much writing can be done during that time. When I am not writing, I used to read a lot of research papers and see how researchers are currently using Deep learning in the genomics field in various domains such as healthcare, agriculture, pharmaceuticals, and so on. Furthermore, as soon as I delivered a chapter, I didn’t wait for the feedback but instead start the next chapter and this allowed me to continue the momentum. I can go on and on, but I would say the main takeaways are incremental progress, focused writing, and a clear thought process.

Q: What’s your take on the technologies discussed in the book? Where do you see these technologies heading in the future?

Upendra: The main technology that was discussed in the book is Deep Learning. It is well known in the industry that Deep learning is revolutionizing every field it was applied to whether it is life science, biotechnology, genomics, and so on. The one exception to this rule is the adoption of Deep learning to different fields varies from one field to another. For example, Deep learning as applied to fields such as Computer Vision and Natural Language Processing has taken off significantly compared to genomics. Genomics has scientifically proven its abilities in the prevention, management, and treatment of disease. Healthcare is the major beneficiary of genomics and because of incredible progress in genomics, the healthcare environment is gradually shifting from conventional treatment methods toward precision medicine. The field of genomics is one of the fastest-growing markets in the world right now and is projected to grow ~100 billion USD by 2028. The global impact of COVID-19 has contributed significantly to this staggering growth and because of that genomics market growth is witnessing a positive demand from all sectors. The recent advances in high-throughput sequencing technologies that generate large-scale data, improvements in algorithms, and the development of next-generation hardware all contributed toward the adoption of Deep learning in genomics. I would imagine Deep Learning would be more routinely applied to genomics in life sciences and biotechnology because of the above reasons and I look forward to a future where this intersection would be talked about at the same level of Deep for Computer Vision or Natural Language Processing.

Q: Why should readers choose this book over others already on the market? How would you differentiate your book from its competition?

Upendra: The simple answer is there is no other book out there that combines Deep learning with genomics. Currently, a lot of resources are available for understanding deep learning but very few resources exist for a proper understanding of deep learning applications for genomics to solve real-world problems. The intersection of Deep learning with genomics has not been explored in life sciences and biotechnology through books. A keyword search of “Deep learning” and “genomics” on google returns 0 books which indicate the gap in the market for this book. The closest I can think of is the Machine learning books that were written to address biological problems in life sciences and biotechnology but not specifically address genomics problems with Deep learning. This book differentiates itself from other books in many ways as it deals with the intersection of deep learning with genomics the two most popular subjects, provides a clear conceptualization of topics that are proven to be important for addressing problems in genomics, including real-world use cases that are routinely been used in industries, and includes topics that help get started with Deep learning for genomics.

Q: What are the key takeaways you want readers to come away with from the book?

Upendra: There are several key takeaways that I want readers to come away with from the book. The main takeaway is that Deep learning is very powerful, easy to implement, and enables one to extract insights from big genomics data. Deep learning is currently being used for genomics applications in both companies and academia successfully. Unlike state-of-the-art technology such as bioinformatics for genomics which relies on rules, the readers will use deep learning algorithms introduced in this book for some of the practical applications of genomics in life sciences and biotechnology industries to transform raw genomics data into valuable knowledge. Going forward, the field of genomics is ready to adopt this new revolution which is deep learning.

Q. What advice would you give to readers learning tech? Do you have any top tips?

Upendra: As with any other technology, the learning curve for Deep Learning will be steep and so it is important to acknowledge that strive towards learning it with incremental steps. One cannot be an expert in any technology overnight, it requires constant practice, getting better at concepts, reading research papers/blogs/subscribing to tech newsletters/following experts in the field, and so on. Finding a project and using that project for learning the new technology is the most widely accepted process for learning the new technology. Amongst all the resources, finding a good book that teaches technology is one of the best ways to learn new technology. For example, if you want to get better at Cloud computing then finding a good book from a reputed publisher and author that teaches cloud computing is way better than googling for relevant materials for cloud computing. The online material is flooded with lots of information but reading a book would allow one to get the most relevant information that one would otherwise find by searching on the internet.

We spend so much of our lives working, so you should enjoy what you do, including the process of learning about it.

Q. Do you have a blog that readers can follow?

Upendra: I am not an active blogger but authoring this book helped me to understand the importance of writing and how it helps to understand concepts much better. So, I decided to spend more time to write more, and I started a blog on my website. Readers can check out this link for more information about it https://upendrak.github.io

Q: Can you share any blogs, websites and forums to help readers gain a holistic view of the tech they are learning?

Upendra: I refer to lots of peer-reviewed research papers to understand a concept (for example, Convolutional Neural Networks in genomics) because they are obviously peer reviewed and so they all went through a rigorous reviewing process and so we can trust them. I follow experts in Machine Learning, Deep Learning, genomics, and Bioinformatics on Twitter and LinkedIn, I read tutorials and blogs on Medium, Towards Data Science, and other websites. So, in summary, no one resource gives you a holistic view of the tech that you are trying to learn but a combination of all of these will get you there eventually.

Q. Do you belong to any tech community groups?

Upendra: I am Carpentries and DataCamp instructor in Data Science.

Q. How would you describe your author’s journey with Packt? Would you recommend Packt to aspiring authors?

Upendra: As I mentioned earlier, my author’s journey with Packt overall is very satisfying even though it was tough towards the end when you are trying to wrap things up. One thing for sure is Packt team will give you all the guidance, help, and direction that you need, and they will be working with you throughout the whole journey, so you don’t feel left alone. That said, book writing is not trivial as it takes so much out of you and many times you feel frustrated, dejected, and disappointed but once you overcome those phases with some positivity, then it is rewarding. I am yet to reap the rewards since the book is not out yet, but I am hoping that this will help me in my future goal of advancing genomics toward the adoption of deep learning. I would recommend Packt to aspiring authors but with the caveats that I mentioned above. But if you are interested in authoring and if you have a clear idea of what you want to teach the readers, then I would highly encourage aspiring authors to author a book.

Q. What are your favorite tech journals? How do you keep yourself up to date on tech?

Upendra: Since my background is not tech, I don’t have read a lot of tech journals but since Biotechnology and Deep Learning have a technology component to them, I read a lot of articles, blogs, and tutorials on Medium and other blogging sites. I am one of the active users on social medium platforms such as Twitter and LinkedIn.

Q. How did you organize, plan, and prioritize your work and write the book?

Upendra: I authored the book in the evenings and weekends (taking care to protect family time at the weekends), and therefore it pays to be organized. When authoring the initial drafts of each chapter, break down the topics or areas in to small chunks that can be tackled in an evening or over a weekend. When working with the editorial team on their review or a technical reviewers comments, again make sure to break down the comments into small sets of changes that can be tackled over the course of evenings or weekends. This way, you can tick off your list as you progress through and continue to feel like progress is being made on the book.

Q. What is that one writing tip that you found most crucial and would like to share with aspiring authors?

Upendra: For aspiring authors, the one tip I want to give is to do your homework during project scheduling and ask for enough time to make sure that the chapters are delivered promptly as scheduled.

Q. Would you like to share your social handles? If so, please share.

Upendra: upendra_35 (Twitter); https://www.linkedin.com/in/upendradevisetty/

You can find Upendra’s book on Amazon by following this link: Please click here

Deep Learning for Genomics is Available on Amazon.com