Special Interview with Frontiers in Bioscience-Landmark Authors Prof. Gregory Warr & Prof. Les Hatton: Interdisciplinary Insights into Transcriptome Structure and Inverse Symmetry
12 May 2026
In this IMR Press interview, we are honored to be joined by two distinguished authors and long‑time research partners, Prof. Gregory Warr and Prof. Les Hatton. They published their article last year (2025, Volume 30, Issue 11) – "The Fine Structure of the Transcriptome: Does It Reflect the Inverse Symmetry of the Genome?" – in Frontiers in Bioscience-Landmark. We are delighted to host this special interview with them. The two first met as undergraduates at King’s College, Cambridge in 1967, where they studied biological sciences and mathematics, respectively. Although their professional paths diverged for several decades, they later reunited around a shared scientific interest: the physical laws that govern both natural and man-made systems.
Gregory Warr
Department of Biochemistry and Molecular Biology, Medical University of South Carolina, Charleston, SC, USA
Interests: biochemistry and molecular biology
Les Hatton
Formerly School of Computer Science and Mathematics, Kingston University, London, UK
Interests: computer science; software system failure analysis
Prof. Warr previously served as Professor of Biochemistry and Molecular Biology at the Medical University of South Carolina and later as a Program Director at the U.S. National Science Foundation. Prof. Hatton spent much of his career in industry and made significant contributions to software engineering and complex systems, receiving the Conrad Schlumberger Award in 1987. In recent years, their interdisciplinary collaboration has brought together biology and information theory to explore universal patterns in living systems.
In this interview, they share insights into their latest research and discuss statistical laws in the transcriptome, complex systems theory, and the future of interdisciplinary science.
1. Could you briefly introduce the central research question of your paper? In addition, could you share your respective roles in this collaboration, highlight the most important conclusions of your study, and tell us how you came together to pursue this kind of interdisciplinary research, coming from biochemistry and computer science, respectively?
Prof. Gregory Warr: About fifteen years ago, Les and I realized we shared an interest in the physical laws governing both the natural and man‑made worlds, especially the pervasive inequality in social and natural systems. Les had been doing research on this and wanted to move it into the biological sciences, because biology is often seen as so complex and specialized that it might be exempt from many natural laws. Therefore, he contacted me, and that is how our collaboration began.
Prof. Les Hatton: It all started around 2008 when I was studying failures in software control systems. By breaking computer programs down into their component pieces, I found that systems written in different languages for different purposes all showed the same statistical behavior. Gregory and I call this “emergent behavior”. A colleague suggested I look at biology – I knew almost nothing about it – so I reached out to Gregory. Since about 2013, we have merged Gregory’s expertise in biochemistry and biological sciences with my skills in computing, mathematics, and information theory. The areas between disciplines are often the most productive in science, and it has been very productive for us. The only difficulty is that journals sometimes struggle to review interdisciplinary papers, but we eventually found excellent journals like yours. After thirteen or fourteen years, we still feel we’ve only scratched the surface.
2. As long-term collaborators, could you share what initially motivated you to explore the statistical properties of the transcriptome and their relationship with genomic inverse symmetry?
Prof. Gregory Warr: Let me step back a bit and talk about the physical principle underlying all our collaborative research – the Conservation of Hartley‑Shannon Information (CoHSI). Les developed the mathematical proof for this principle. It explains the consistent and widespread emergent properties in biological and natural systems, particularly power‑law distributions. We were interested in how this principle interacts with biological systems. CoHSI predicts an equilibrium and presents the opportunity to study how and why the pressures of evolution on living systems can perturb that equilibrium. By studying those perturbations, we gain insight into what is going on inside the systems. The transcriptome is at the heart of life – it is where DNA information is converted into proteins and other essential components – so it was a natural topic for us. Our research is an iterative process: Les identifies mathematical questions and devises computational methods, and I, as the biologist, provide the biological background and help design and interpret the experiments.
3. Gregory Warr, one of your most striking findings is that stop codons are outliers in the reading frame analysis. As someone who understands translation intimately, what did you think when you first saw that plot?
Prof. Gregory Warr: The two most rewarding things in a scientific research career are, first, the opportunity to see things nobody else has seen before, and second, to see how beautifully and logically pieces of a puzzle fall into place. Les and I often use different independent approaches; if the results fit together, we feel we have a handle on the correct answer. In this study, we compared codon frequencies in the correct reading frame versus an incorrect reading frame. For non‑coding RNA, the frequencies are essentially identical regardless of the reading frame. But for coding sequences, the result is very different – the frequencies of all codons differ greatly between the correct and incorrect frames, and stop codons in particular show about a ten‑fold difference. This makes perfect sense, because stop codons are the most powerful signals in the genetic code – they bring translation to an absolute halt. Seeing that result was very satisfying.
4. Les Hatton, you have long been engaged in research on software failures and vulnerabilities. Could you share what intrinsic connections exist between this work on the statistical properties of the transcriptome and your previous research?
Prof. Les Hatton: I spent about 25 years studying how computer systems fail. I believe most scientific software is corrupted by unquantifiable errors. I started looking at systems as a whole and then went down into their components, much like breaking a book into chapters, paragraphs, sentences, and words. I wanted to find patterns that might correlate with failure. Although I failed to find a sensible correlation, I discovered that the length of components in any software package always obeys the same relation – regardless of language, author, or purpose. This smelled of a conservation principle. It took me about four months to prove what we call the asymptotic version: the length of software components is independent of the meaning of the tokens, depending only on the fact that they can be distinguished. I published that in 2014. A colleague suggested I look at biology. I contacted Gregory, and we soon found that the theory that predicts protein lengths is also independent of the meaning of their amino acid constituents. In 2015, we predicted that within five years we would see proteins longer than 40,000 amino acids; five years later, the first such protein was discovered. This clearly shows that one of the fundamental properties of life – protein length – has nothing to do with natural selection. Similarly, Zipf’s law for word frequencies arises from the same principle. We have since validated the principle in finance and many other areas. Therefore, the merging of our two disciplines felt very natural.
5. Generative AI and large language models are currently deeply empowering bioinformatics research and are widely used in omics data mining, biological sequence feature analysis, and other directions. Combined with the transcriptome k-mer analysis work of this study, how do you two think we can combine large language models with traditional statistical analysis methods to mine the sequence structure laws of the genome/transcriptome more efficiently and accurately? What interdisciplinary technical difficulties still need to be resolved in order to develop such AI tools to analyze biological sequences?
Les Hatton: Combining large language models with statistical methods is still in its infancy. Whatever statistical method you try to teach them, they currently make far too many plausible mistakes. But things are rapidly changing. Regarding the areas we’ve been working on, CoHSI is essentially a statistical method – it’s based on statistical mechanics, closely related to the Maxwell‑Boltzmann law. Therefore, as far as we succeed in teaching generative AI to understand logic and statistical methods, CoHSI is just another statistical method that they need to be aware of. Once AI has been taught to handle logical methods, incorporating CoHSI should not be a major additional step. They should merge seamlessly, just as with any other statistical method.
6. The two of you are from the fields of biochemistry and molecular biology, computer science, and mathematics, respectively, and this interdisciplinary cooperation has achieved excellent research results. For the development of interdisciplinary research in the future, what new cooperation directions and ideas do you have?
Prof. Gregory Warr: I think our approach has been extremely successful in analyzing discrete systems. We can explain why they reach equilibrium and how specific pressures cause perturbations in that equilibrium. But what we would really like to do is to predict where these discrete systems are going – their future course. Unfortunately, based on the mathematics alone, we cannot predict the future course of biological evolution. We’re not there yet.
Prof. Les Hatton: I looked at the notes I made for this interview, and next to this question, I wrote “too many for a lifetime.” Every time we do something, it uncovers two more interesting subjects. In the past fifteen years, this project has produced 35,000 figures, most of which we’ve never used. It’s a very fertile area. The genetics of discrete systems seems to offer so many other feasible and insightful directions. For example, as we speak, Artemis is on its way back from the moon – it might interest you to know that the widths of craters on the moon follow the same distribution as the genome. Two very different discrete systems.
7. Why did the two of you choose to submit this important interdisciplinary research result to Frontiers in Bioscience-Landmark (FBL)? What characteristics of our journal attracted you?
Prof. Gregory Warr: I think we were flattered that the editors of FBL contacted us based on their review of the literature and asked if we would be interested in submitting our work to them. We looked at the journal and were impressed by the wide range of topics it considers and publishes. It was clear to us that there wasn’t an agenda – no “diamond science” as one might call it. The journal is open to science in many different areas without prejudice. So we decided to submit this paper, which was just reaching fruition. We’re very happy we did.
8. What was your overall experience during the entire process of submission, peer review, revision, and final publication of this review? What details in the manuscript processing process of FBL left a deep impression on you?
Prof. Gregory Warr: Let me talk about the sociology of science. Many fields have specialized journals, and certain schools of thought can come to dominate how data are interpreted and what is considered publishable. Les and I have encountered a great deal of this during our collaborative research. It can be very frustrating when you have what you think is a new and challenging way of looking at biological systems – especially when it questions the role of natural selection in generating these patterns. So we were very happy that Frontiers in Bioscience‑Landmark reviewed our work very carefully and very objectively, on its own terms, rather than trying to fit it into a preconceived framework. The review was good and objective, aimed at making the paper the best it could be. I have extensive experience editing journals, and this was one of the best experiences we have had as authors submitting manuscripts for review.
Professors Warr and Hatton have shared with us their interdisciplinary journey from software engineering and information theory to biology, explaining in depth how the CoHSI principle reveals hidden statistical laws in the transcriptome, along with exciting findings such as stop codon anomalies and predictions of protein length. We thank Professors Warr and Hatton for accepting this special interview with Frontiers in Bioscience‑Landmark and for their sincere and insightful sharing. Our journal remains committed to publishing rigorous, innovative, and interdisciplinary scientific results, and we deeply value scholars like you who dare to break down disciplinary boundaries. We look forward to witnessing more of your remarkable research in the future. Thank you both for submitting this important research to our Journal.
Article Details: The Fine Structure of the Transcriptome: Does It Reflect the Inverse Symmetry of the Genome?
Journal Homepage: Frontiers in Bioscience-Landmark