Scientists at Trinity College, Dublin, Ireland have discovered that using the natural algorithm of evolution by natural selection is the best way to improve the efficiency of enzymes in terms of the products of the processes they catalyse.
This begs an embarrassing question for creationists: why can enzymes produced by a perfect designer be improved upon?
The answer of course, is that they weren't designed by a perfect designer; they were designed by a utilitarian evolutionary process in which whatever is better than what went before will be retained and the only reason to improve beyond the sub-optimal that works well enough, is to have some other selectors driving it towards ever greater fitness. There is no 'ideal' nor any process for comparing what is against this idea and trying harder, as there would be if there was an intelligence guiding the process.
In general terms, what is a genetic algorithm? A genetic algorithm (GA) is a search heuristic and optimization technique inspired by the principles of natural evolution. It is used to find approximate solutions to complex problems by iteratively improving a population of candidate solutions. Here's an overview of the key components and processes involved in a genetic algorithm:It is in the latter function that science comes in with a predetermined measure of optimal performance against which to compare the current performance and running the process again if it doesn't measure up.
Key Components
- Population:
- A set of candidate solutions to the problem. Each candidate, often called an individual or chromosome, represents a potential solution encoded in a suitable format, typically as a string of bits, numbers, or symbols.
- Genes:
- Basic units of the chromosome that represent specific traits or parameters of the solution. In biological terms, genes are analogous to specific segments of DNA.
- Fitness Function:
- A function that evaluates and assigns a fitness score to each individual in the population based on how well it solves the problem or meets the desired criteria.
- Genetic Operators:
- Selection: Process of choosing individuals from the current population to create offspring for the next generation, based on their fitness scores.
- Crossover (Recombination): Combining parts of two parent chromosomes to create one or more offspring with traits from both parents.
- Mutation: Introducing random changes to individual genes to maintain genetic diversity within the population and explore new solutions.
Process
- Initialization:
- Generate an initial population of individuals randomly or using some heuristic method.
- Evaluation:
- Calculate the fitness of each individual in the population using the fitness function.
- Selection:
- Select individuals based on their fitness scores to serve as parents for the next generation. Common methods include roulette wheel selection, tournament selection, and rank selection.
- Crossover:
- Pair selected parents and exchange segments of their chromosomes to produce offspring. Various crossover methods exist, such as single-point crossover, multi-point crossover, and uniform crossover.
- Mutation:
- Apply random changes to the offspring's genes with a certain probability (mutation rate). This step helps introduce new genetic material into the population.
- Replacement:
- Form a new generation by replacing some or all of the old population with the new offspring. Strategies include generational replacement (replacing the entire population) or steady-state replacement (replacing only a few individuals).
- Iteration:
- Repeat the evaluation, selection, crossover, mutation, and replacement steps for multiple generations until a stopping criterion is met. The criterion could be a maximum number of generations, a satisfactory fitness level, or convergence of the population.
Applications
Genetic algorithms are widely used in various fields, including:
- Optimization Problems: Finding optimal or near-optimal solutions for complex functions, resource allocation, scheduling, and design.
- Machine Learning: Feature selection, hyperparameter tuning, and evolving neural network architectures.
- Engineering: Designing efficient structures, control systems, and electronic circuits.
- Bioinformatics: Protein folding, genetic sequencing, and drug design.
Advantages and Limitations
Advantages:
- Capable of handling complex, multi-dimensional, and non-linear problems.
- Do not require gradient information or derivative calculations.
- Can escape local optima and explore a broad search space.
Limitations:
- Computationally expensive, especially for large populations and many generations.
- May converge prematurely to suboptimal solutions if not properly tuned.
- Performance depends on the choice of genetic operators, parameters, and fitness function.
Genetic algorithms provide a robust and flexible framework for solving challenging problems through evolutionary principles, leveraging the power of natural selection and genetic variation.
The problem with trying to improve on the performance of an enzyme is in the vast number of different permutations that are possible with every amino acid in the protein chain have 20 possible alternatives, most of which are going to be detrimental or at best neutral, so what the Trinity College team have done is to look at the evolutionary history of an enzyme and see where changes made an improvement and build on those.
The enzyme they used to test the principle was TEM-1 β-lactamase and what the team were looking for were improved thermostability (in other words it was stable over a bigger range of temperatures) and increased activity across a range of substrates.
Their research is explained in a Trinity College news release and in an open access paper in the journal Nature Communications:
Scientists devise algorithm to engineer improved enzymes
The scientists have devised an algorithm, which takes into account an enzyme’s evolutionary history, to flag where mutations could be introduced with a high likelihood of delivering functional improvements.
Their work – published today in leading journal Nature Communications – could have significant, wide-ranging impacts across a suite of industries, from food production to human health.
Enzymes are central to life and key to developing innovative drugs and tools to address society’s challenges. They have evolved over billions of years through changes in the amino acid sequence that underpins their 3D structure. Like beads on a string, each enzyme is composed of a sequence of several hundred amino acids that encodes its 3D shape.
With one of 20 amino acid ‘beads’ possible at each position, there is enormous sequence diversity possible in nature. Upon formation of their 3D shape, enzymes carry out a specific function such as digesting our dietary proteins, converting chemical energy into force in our muscles, and destroying bacteria or viruses that invade cells. If you change the sequence, you can disrupt the 3D shape, and that typically changes the functionality of the enzyme, sometimes rendering it completely ineffective.
Finding ways to improve the activity of enzymes would be hugely beneficial to many industrial applications and, using modern tools in molecular biology, it is simple and cost-efficient to engineer changes in the amino acid sequences to facilitate improvements in their performance. However, randomly introducing as little as three or four changes to the sequence can lead to a dramatic loss of their activity.
Here, the scientists report a promising new strategy to rationally engineer an enzyme called “beta-lactamase”. Instead of introducing random mutations in a scattergun approach, researchers at the Broad Institute and Harvard Medical School developed an algorithm that takes into account the evolutionary history of the enzyme.
At the heart of this new algorithm is a scoring function that exploits thousands of sequences of beta-lactamase from many diverse organisms. Instead of a few random changes, up to 84 mutations over a sequence of 280 were generated to enhance functional performance. And strikingly, the newly designed enzymes had both improved activity and stability at higher temperatures.
Associate Professor Dr Amir Khan, co-author
School of Biochemistry and Immunology
Trinity College, Dublin, Ireland.
Eve Napier, a second-year PhD student at Trinity, determined the 3D experimental structure of a newly designed beta-lactamase, using a method called X-ray crystallography. Her 3D map revealed that despite changes to 30% of the amino acids, the enzyme had an identical structure to the wild-type beta-lactamase. It also revealed how coordinated changes in amino acids, introduced simultaneously, can efficiently stabilise the 3D structure – in contrast to individual changes that typically impair the enzyme structure.
Overall, these studies reveal that proteins can be engineered for improved activity by dramatic ‘jumps’ into new sequence space. The work has wide ranging applications in industry, in processes that require enzymes for food production, plastic-degrading enzymes, and those relevant to human health and disease, so we are quite excited for the future possibilities.
Eve Napier, co-author
School of Biochemistry and Immunology
Trinity College, Dublin, Ireland.
Abstract
A major challenge in protein design is to augment existing functional proteins with multiple property enhancements. Altering several properties likely necessitates numerous primary sequence changes, and novel methods are needed to accurately predict combinations of mutations that maintain or enhance function. Models of sequence co-variation (e.g., EVcouplings), which leverage extensive information about various protein properties and activities from homologous protein sequences, have proven effective for many applications including structure determination and mutation effect prediction. We apply EVcouplings to computationally design variants of the model protein TEM-1 β-lactamase. Nearly all the 14 experimentally characterized designs were functional, including one with 84 mutations from the nearest natural homolog. The designs also had large increases in thermostability, increased activity on multiple substrates, and nearly identical structure to the wild type enzyme. This study highlights the efficacy of evolutionary models in guiding large sequence alterations to generate functional diversity for protein design applications.
Introduction
As proteins become increasingly useful across a range of fields including medicine and industry, there is a growing need for designed proteins with optimized characteristics, such as elevated thermostability, higher binding affinity, or increased catalytic activity. Natural proteins are often used as starting points for the development of useful proteins, which can then be engineered as high-performance, task-specific tools. However, efficiently mutating enzymes to yield optimized variants is exceedingly difficult, and randomly mutating enzymes almost always leads to loss of performance, which decreases considerably with every additional mutation1. Information-based ‘rational’ engineering can avoid performance loss, but is generally limited to a very small number of sequence changes. One approach to protein engineering, directed evolution, makes use of iterative rounds of mutagenesis followed by selection to optimize a specific property like activity or thermostability. However, increased random mutation count overwhelmingly negatively impacts fitness1, limiting the number of amino acid changes that can be introduced while still maintaining a reasonable number of functional variants. This stepwise incremental selection strategy is often effective at finding sequences with improved properties with a limited number of mutations. The introduction of many simultaneous changes to a protein’s primary sequence is likely required to diversify and optimize multiple desirable properties, and new methods that enable such large changes in primary sequence are needed. Computational design strategies2,3,4,5,6,7, which account for the complexity of how each mutated residue interacts with all other residues, are likely required to maintain function when introducing more than a handful of mutations.
An evolution-informed computational protein design strategy may provide a means to generate many changes in primary sequence, enabling the exploration of diverse structural and functional properties. Evolutionary models that account for complex selective conditions over millions of years by learning meaningful constraints on function from related sets of homologous protein sequences5,8,9,10 have been shown to recapitulate core aspects of protein biology, such as 3D structure8,11,12, protein stability10,13,14, conformational state15,16,17 and the effects of mutations on protein fitness2,10,13,18,19,20. Some of these models have been used for protein design, generating mutated proteins from a wild type scaffold that maintain structure and/or function2,5,6,7,21,22,23. Evolutionary couplings (EVcouplings) models are a specific instance of evolutionary models based on residue site- and pairwise dependencies in natural sequence variation8,10,24. These models are unsupervised, inferring sequence constraints characteristic of a functional space and quantifying fitness differences between variants without experimentally measured phenotype labels. To make use of these discriminative models for protein design, a sampling algorithm is used to iteratively generate variant sequences that are chosen to optimize a fitness function.
The TEM-1 β-lactamase model system has been extensively used to study protein evolution7,9,25,26,27,28. β-lactamases are a class of enzymes that are produced by bacteria in order to provide resistance to bactericidal β-lactam antibiotics through hydrolysis of their core β-lactam ring. Many bench biologists are familiar with the use of TEM-1 β-lactamase as a marker for successful transformation, in which selection of functional TEM-1-containing plasmids is as simple as growth in a β-lactam antibiotic like ampicillin. Due to experimental tractability, many publications report the effects of mutations on TEM-1 function and stability, including several studies using deep mutational scans25,27,29. Other studies have described the exponential decrease in TEM-1 function when subjected to multiple mutations, with the cumulative effect of 10 random mutations completely abrogating enzyme activity1.
In this work we investigate whether evolutionary models of sequence co-variation can be used to design enzyme variants that contain many changes to the target sequence while maintaining function. In addition, we test whether making large jumps in the primary amino acid sequence can lead to augmented protein properties such as increased thermostability, increased activity and broadening of available substrates, and investigate the implications of these mutations on protein 3D structure.
Fram, B., Su, Y., Truebridge, I. et al.
Simultaneous enhancement of multiple functional properties using evolution-informed protein design.
Nat Commun 15, 5141 (2024). https://doi.org/10.1038/s41467-024-49119-x
Copyright: © 2024 The authors.
Published by Springer Nature Ltd. Open access.
Reprinted under a Creative Commons Attribution 4.0 International license (CC BY 4.0)
Creationists who don't see this sort of research as a problem, probably haven't understood it or grasped the significance of it. What they need to consider is:
- How can something designed by a perfect designer be improved upon?
- Why is a genetic algorithm such a good tool for finding novel ways to improve a system?
- If intelligently devising an ideal outcome and testing against that ideal produces a better design, why does what we see in nature look like something produced without a concept of an ideal outcome, resulting in something that can be improved, using intelligence?
What Makes You So Special? From The Big Bang To You
How did you come to be here, now? This books takes you from the Big Bang to the evolution of modern humans and the history of human cultures, showing that science is an adventure of discovery and a source of limitless wonder, giving us richer and more rewarding appreciation of the phenomenal privilege of merely being alive and able to begin to understand it all.
Available in Hardcover, Paperback or ebook for Kindle
Ten Reasons To Lose Faith: And Why You Are Better Off Without It
This book explains why faith is a fallacy and serves no useful purpose other than providing an excuse for pretending to know things that are unknown. It also explains how losing faith liberates former sufferers from fear, delusion and the control of others, freeing them to see the world in a different light, to recognise the injustices that religions cause and to accept people for who they are, not which group they happened to be born in. A society based on atheist, Humanist principles would be a less divided, more inclusive, more peaceful society and one more appreciative of the one opportunity that life gives us to enjoy and wonder at the world we live in.
Available in Hardcover, Paperback or ebook for Kindle
No comments:
Post a Comment
Obscene, threatening or obnoxious messages, preaching, abuse and spam will be removed, as will anything by known Internet trolls and stalkers, by known sock-puppet accounts and anything not connected with the post,
A claim made without evidence can be dismissed without evidence. Remember: your opinion is not an established fact unless corroborated.