Dec 10, 2023·edited Dec 10, 2023Liked by Erika Alden DeBenedictis, Asimov Press
As a researcher who does synthetic developmental biology, I'd add another example of a predictive model that would be very useful: can we predict the developmental effects of a biological perturbation (such as expressing a protein in a certain cell type, mutating a transcription factor binding site, or inhibiting a kinase)? Even just predicting the effects on the cellular transcriptome and/or proteome would be very useful.
Also I don't think this statement is accurate: "Codon optimizers have effectively solved the DNA → RNA part of protein expression"
Codon optimizers don't affect DNA->RNA (transcription), they affect RNA->protein (translation).
The statement "In the coming decades, we may actually see predictive models that help biologists express any protein, with any function, in any organism" reflects an important distinction between the fields of physics and biology.
Systems Biology Perspective: In biology, especially in fields like systems biology, there's a growing recognition that understanding biological processes requires a holistic view of living systems. This involves considering not just individual proteins but also the interactions between various molecular components (proteins, RNA, DNA, metabolites) and the influence of the cellular and environmental context. Predictive models in biology, therefore, need to encompass the complexity of these systems, acknowledging that biological functions often emerge from intricate networks of interactions rather than isolated components.
Challenges in Predictive Modeling for Biology: The idea of predicting the expression of "any protein, with any function, in any organism" is an ambitious goal. However, it simplifies the complexity of biological systems. Biological systems are not just about expressing a protein but also about how that protein interacts within a cell, how it's modified, how it degrades, and how it affects and is affected by the organism's environment. Moreover, biological systems exhibit a high degree of variability and adaptability, which can be challenging to capture in predictive models.
Interdisciplinary Approaches: The statement reflects a more reductionist approach typical in physics, where systems are often simplified to understand fundamental principles. While this approach has its advantages, biology often requires a more integrative approach that considers multiple levels of organization from molecules to cells to organisms. Therefore, successful predictive models in biology often require collaboration across disciplines, integrating insights from biochemistry, genetics, computational biology, ecology, and other areas.
Potential of Integrative Models: The development of integrative models that consider entire systems rather than just individual components could lead to more accurate predictions and a deeper understanding of biological phenomena. This includes understanding the dynamics of biological networks, gene regulation, metabolic pathways, and ecological interactions, which are crucial for advancing our knowledge in fields like medicine, agriculture, and environmental science.
In conclusion, while the ambition to develop predictive models in biology is commendable, it's essential to recognize the inherent complexity and interconnectedness of biological systems. Advances in the field will likely come from approaches that integrate multiple levels of biological organization, rather than focusing solely on individual components like proteins.
Kudos for the excellent and thought-provoking essay. Two comments: 1. The statement that in physics "great discoveries today seem to stem mostly from large programs with multi-billion dollar price tags." needs a reality check from some other metric than the successful celebrity of the large programs, e.g. metrics like opinions of physicists, Noble prizes to small or large projects, publication citations, patents and conversion of discoveries into commercial or public goods could be relevant. This might wind up supporting your position but might not. 2. The easiest way to improve large-scale progress in biology would be a more rigorous legislative and funder insistence (with draconian penalties for failure) on putting data into public databases. You will be familiar with both widespread anecdotal and published studies reporting the widespread failure of bio researchers to share their data. The wonderful success of the opposite behavior (e.g. UK Biobank, GEO, etc) could easily be amplified if failure to share meant enforced ineligibility for future funding.
As a researcher who does synthetic developmental biology, I'd add another example of a predictive model that would be very useful: can we predict the developmental effects of a biological perturbation (such as expressing a protein in a certain cell type, mutating a transcription factor binding site, or inhibiting a kinase)? Even just predicting the effects on the cellular transcriptome and/or proteome would be very useful.
Also I don't think this statement is accurate: "Codon optimizers have effectively solved the DNA → RNA part of protein expression"
Codon optimizers don't affect DNA->RNA (transcription), they affect RNA->protein (translation).
The statement "In the coming decades, we may actually see predictive models that help biologists express any protein, with any function, in any organism" reflects an important distinction between the fields of physics and biology.
Systems Biology Perspective: In biology, especially in fields like systems biology, there's a growing recognition that understanding biological processes requires a holistic view of living systems. This involves considering not just individual proteins but also the interactions between various molecular components (proteins, RNA, DNA, metabolites) and the influence of the cellular and environmental context. Predictive models in biology, therefore, need to encompass the complexity of these systems, acknowledging that biological functions often emerge from intricate networks of interactions rather than isolated components.
Challenges in Predictive Modeling for Biology: The idea of predicting the expression of "any protein, with any function, in any organism" is an ambitious goal. However, it simplifies the complexity of biological systems. Biological systems are not just about expressing a protein but also about how that protein interacts within a cell, how it's modified, how it degrades, and how it affects and is affected by the organism's environment. Moreover, biological systems exhibit a high degree of variability and adaptability, which can be challenging to capture in predictive models.
Interdisciplinary Approaches: The statement reflects a more reductionist approach typical in physics, where systems are often simplified to understand fundamental principles. While this approach has its advantages, biology often requires a more integrative approach that considers multiple levels of organization from molecules to cells to organisms. Therefore, successful predictive models in biology often require collaboration across disciplines, integrating insights from biochemistry, genetics, computational biology, ecology, and other areas.
Potential of Integrative Models: The development of integrative models that consider entire systems rather than just individual components could lead to more accurate predictions and a deeper understanding of biological phenomena. This includes understanding the dynamics of biological networks, gene regulation, metabolic pathways, and ecological interactions, which are crucial for advancing our knowledge in fields like medicine, agriculture, and environmental science.
In conclusion, while the ambition to develop predictive models in biology is commendable, it's essential to recognize the inherent complexity and interconnectedness of biological systems. Advances in the field will likely come from approaches that integrate multiple levels of biological organization, rather than focusing solely on individual components like proteins.
Kudos for the excellent and thought-provoking essay. Two comments: 1. The statement that in physics "great discoveries today seem to stem mostly from large programs with multi-billion dollar price tags." needs a reality check from some other metric than the successful celebrity of the large programs, e.g. metrics like opinions of physicists, Noble prizes to small or large projects, publication citations, patents and conversion of discoveries into commercial or public goods could be relevant. This might wind up supporting your position but might not. 2. The easiest way to improve large-scale progress in biology would be a more rigorous legislative and funder insistence (with draconian penalties for failure) on putting data into public databases. You will be familiar with both widespread anecdotal and published studies reporting the widespread failure of bio researchers to share their data. The wonderful success of the opposite behavior (e.g. UK Biobank, GEO, etc) could easily be amplified if failure to share meant enforced ineligibility for future funding.