AI Supercharges Synthetic Biology
Synthetic biology and AI: Revolutionizing life as we know it. Explore gene editing, bioprinting organs, creating synthetic life, and integrating electronics into biology.
Chapter 21 of the book “Machines of Tomorrow: From AI Origins to Superintelligence & Posthumanity.” How AI Will Shape Our World.
Buy “Machines of Tomorrow” on Amazon
Join our substack
Review it on Amazon or Goodreads
“If you get a personal genome, you should be able to get personal cell lines, stem cell derived from your adult tissues, that allow you to bring together synthetic biology and the sequencing so that you can repair parts of your body as you age or repair things that were inherited disorders.”
George M Church
Synthetic Biologist
Interview in ThinkBig.com [Church]
2017
Synthetic biology (SynBio) is an interdisciplinary field that integrates principles from biology, engineering, and, most importantly, AI to design and construct new biological systems or redesign existing ones for practical applications. It employs engineering principles on living organisms and allows scientists to manipulate genetic material and other biological components. At its core, synthetic biology seeks to treat biological components as interchangeable building blocks, similar to the way engineers approach computer design with separate individual electronic circuits.
Synthetic biology techniques have been applied to engineer microorganisms capable of converting renewable resources like plant biomass or algae into biofuels such as ethanol or biodiesel that can be used as a replacement for fossil fuels. Synthetic biology has also been used to design engineered organisms, like bacteria or plants, that can be programmed to absorb specific pollutants, thereby facilitating the cleanup of contaminated sites and contributing to environmental conservation.
The pharmaceutical industry is a chief beneficiary of technological progress, using synthetic biology advances to design microorganisms capable of synthesizing complex pharmaceutical compounds and drugs. This technique accelerates drug production processes in a cost-effective way and could hold the potential to increase supply and accessibility to essential medications. Synthetic biology also plays a pivotal role in designing and synthesizing proteins, such as enzymes or antibodies, with enhanced functions that would open avenues for applications in medicine and industrial processes.
In addition, synthetic biology is venturing into the ambitious realms of creating artificial organs and even supporting and creating novel life forms. By combining biological components with synthetic materials, scientists aim to engineer functional organs for transplantation, addressing the shortage of donor organs and improving the success rates of transplantation procedures. Likewise, scientists are exploring the application of synthetic biology to design and produce living organisms with unique capabilities not found in natural organisms. These synthetic organisms could be engineered for specific industrial, medical, or environmental purposes. Today, we already have examples of non-naturally occurring organisms being fully bio-engineered, so the direction in which the discipline is hurdling becomes quite clear.
For example, there are synthetic yeast strains designed to produce specific flavors and aromas, leading to innovative beer and wine products. Companies like Ginkgo Bioworks have been active in this area. Synthetic algae have been engineered for biofuel, with players like Synthetic Genomics, in collaboration with ExxonMobil, having worked on genetically modifying algae strains to enhance their oil-producing capabilities. Likewise, synthetic bacteria have been engineered for drug production, with companies like Genentech and Novartis having engineered bacteria to produce pharmaceuticals such as insulin and HGH (Human Growth Hormone.) Synthetic bacteria have also been used for environmental cleanup, with several, including Bioremediation Services, having developed them to break down pollutants in soil and water. Finally, companies such as Pivot Bio have engineered synthetic microbes to act as biofertilizers, helping improve nutrient uptake in plants and reduce the need for traditional chemical fertilizers.
The following pages will present the potential for modifying the biological nature of human beings through synthetic biology when AI is applied. We see the impact of synthetic biology and AI solving the lingering question of “how” interlacing will mechanically occur.
Link to the book: Machines of Tomorrow
The Science Fiction of Synthetic Biology
Synthetic Biology is arguably one of the most revolutionary applications of AI. It has the unmitigated potential to transform the essence of human beings and our present ecosystems by altering their biological fundamentals, allowing people to live longer, resist diseases, and acquire superhuman abilities with specifically designed organoids. Moreover, integrating electronic or machine-learning elements into human biological circuits opens a solid avenue for human-AI interlacing and, in our view, solves a concrete “how to” problem that is necessary for interlacing to occur. Furthermore, synthetic biology holds the key to creating new forms of life, some of which would never have developed.
To explore the implications of synthetic biology in humanity, we can turn to two Science Fiction works for insight: the film “Gattaca” directed by Andrew Niccol in 1997, and the novel “The Windup Girl” written by Paolo Bacigalupi in 2009 [Niccol] [Bacigalupi]. While Ridley Scott’s 1982 film “Blade Runner” presents Replicants, the most iconic example of biologically engineered entities in Science Fiction, we have chosen “The Windup Girl” instead because it presents a more carefully developed and multifaceted perspective of this intriguing possibility.
These curated Science Fiction works offer two unique and thought-provoking perspectives on the potential consequences of manipulating life at its fundamental level. Gattaca’s primary focus lies in harnessing synthetic biology to enhance the human race, whereas “The Windup Girl” utilizes synthetic biology to craft a subservient class of individuals. “Gattaca” paints a dystopian vision of a future where genetic engineering has given rise to an overtly rigid and stratified society. Individuals are sorted into a hierarchy of classes based on their genetic makeup, with “valids” who are genetically enhanced and “invalids” who are deemed genetically inferior. This genetic discrimination permeates all aspects of life, with the invalid protagonist forced to assume a valid identity to pursue his dream of becoming an astronaut. The film also explores the themes of identity and destiny as the protagonist grapples with societal norms and expectations, ultimately questioning the very essence of society and the human being.
Synthetic biology’s dedicated ability to manipulate genetic information forces us to ponder the boundaries of human identity and the ethics of designing and selecting genes. As we have pointed out earlier, we believe that the economic costs, timing, and selection factors associated with any element of interlacing will inevitably mean that it happens asymmetrically within society, with some being left behind either by self-selection or fiat. One of the striking similarities between “Gattaca” and the real-world implications of synthetic biology is the likelihood of a genetic divide within society. As synthetic biology advances, there is a high potential for a division between those who can afford genetic enhancements and those who cannot, at least in its early phases, enabling the possibility that those who are first may stop the process before enabling those who come second, further perpetuating inequality in a way that self-compounds. Human nature reflects in consumer transactions — for example, buying a Cadillac or BMW as a functional piece of engineered machinery is a social status symbol that not everyone can have. Synthetic biology products will be a far more absolute and far-reaching reflection.
We note that eugenics, along with the moral and ethical polemics that accompany it, are embedded in any discussion of synthetic biology. The technology makes it not only easy to understand and modify human genes to even subtle differences but also to obviate those differences and their genetic factors. Furthermore, it facilitates the mathematical ranking of capabilities and outcomes, thus determining the relative value of the genetics involved. The mathematical reality is that under a specified set of conditions, some genetic material is more valuable than others. This might sound Machiavellian on the surface because AI will not necessarily have ethical values embedded in it. This omnipresent element of synthetic biology returns us briefly to a central tenet of this book, namely that it is up to us, now, not later, to set the grounding rules for responsible AI in order to avoid a purely dystopian future.
In “The Windup Girl,” the focus shifts to synthetic biology’s impact on the environment and biodiversity. “The Windup Girl” is set in 23rd-century Thailand in a post-oil world. The novel envisions a future where synthetic biology is central to food production and energy generation. Genetically engineered organisms, known as “New People,” are designed to serve various purposes, including labor and entertainment. Soon, in the novel, the lines between artificial and natural life start to be blurred, leading to innumerable questions and contradictions. We discuss the impact of AI in general on irrevocably distorting the line between truth and falsehood in Chapter 23.
One of the critical differences between “The Windup Girl” and “Gattaca” lies in the scope of genetic manipulation. In “The Windup Girl,” synthetic biology extends beyond human genetic engineering to encompass the entire ecosystem. This raises ethical concerns about manipulating the environment and biodiversity for economic or other gain, with potentially unwanted ecological consequences. In the novel, corporations exploit synthetic biology for profit, resulting in a loss of agricultural diversity and dependence on engineered crops.
The novel also delves into the intersection of spirituality and synthetic biology as the engineered protagonist, a member of the “New People,” prompts discussions about the nature of her soul. Characters in the story ponder whether these synthetic people possess spirituality or consciousness, highlighting the interplay between religion and the ethical considerations of creating life through genetic manipulation, a topic we also cover later in Chapter 26.
Synthetic biology has the potential to improve our interactions with nature, including our biology, environment, and food (as suggested by these two novels). The other side of the technology coin is far more ominous, with implications of eugenics and mankind’s long and troublesome history with the concept of genetic differences translating to value differences among humans or unnatural ecosystem manipulation for profit. Simply put, those who control this technology at scale are poised to gain substantial and perhaps absolute control over others.
In another vein, for those familiar with the Captain America character created by Timely (later Marvel) Comics in 1941, the so-called “super soldier serum” that transformed sickly Steve Rogers into the greatest fighting machine of his era points us to both expect military applications of synthetic biology and to foresee constant subterfuge, escalation and machination to control its advance. We have noted throughout the book the key role of DARPA at various points in the evolution of AI and note that the Chinese military is actively involved in the development of synthetic biology and the collection of Western-generated data sets on population biology, a topic we discuss in Chapter 24.
Who is to say if a real-life Captain America would seek to preserve our values or not? There are individuals with psychographic profiles, like Steve Rogers, and those who obtain quite different motives. The huge power embedded in synthetic biology makes us either the masters of creation or the masters of disequilibrium. We will delve into this in Chapter 26 when discussing Superintelligence. In the wake of geopolitical brinkmanship over AI technology, perhaps Captain America is needed regardless.
Understanding the Basics of Biology
Before delving into the application of AI in synthetic biology, it is essential to provide an introduction to the fundamentals of biology and the complexities of life processes. Synthetic biology focuses on modifying or synthesizing these processes.
Life as we know it is a wondrous phenomenon that is fundamentally built upon the functioning of cells. These microscopic units are the fundamental components of all living beings, whether single-celled microorganisms or complex multicellular organisms like humans.
The nucleus at the core of every cell is a crucial component housing genetic information. Within the nucleus resides DNA (Deoxyribonucleic Acid), which carries the genetic instructions to develop, function, and reproduce all living organisms. DNA is a remarkable molecule with a double-helix structure formed by a sequence of molecules called nucleotides. These nucleotides are the basic building blocks of DNA. In nature, there are only four types of nitrogenous bases: adenine (A), thymine (T), cytosine ©, and guanine (G). The nucleotides are like the letters of the language in which life is encoded. The sequence of these bases ultimately encodes the genetic codes that determine an organism’s characteristics and functions.
Genes are specific DNA sequences that carry the instructions for synthesizing proteins. Each gene has a particular sequence of bases that serves as a code, directing the cell to create a specific protein. Proteins are complex and large molecules formed by chains of basic units called amino acids. Each sequence of bases in the DNA corresponds to a specific sequence of amino acids in the protein. RNA (Ribonucleic Acid) plays an intermediary role in translating genetic information from DNA into proteins.
There are many kinds of proteins. Among them, enzymes are particularly important because they serve as vital catalysts for biochemical reactions, aid in cellular communication, and contribute to the structural integrity and functionality of cells and organisms.
Synthetic Biology and Artificial Intelligence
Synthetic biology deals with the highly complex molecular structures of proteins and genes. While it is theoretically possible to design these molecules manually or with simple computer software, AI is indispensable to do it at scale. AI algorithms can analyze vast datasets, predict potential outcomes of genetic configurations and modifications, and optimize the design of biological circuits for specific functions. They can also assist in identifying patterns and correlations within biological data, facilitating the discovery of novel genetic combinations that may yield desired traits. The incorporation of AI into synthetic biology not only accelerates design processes but also improves the precision of crafting synthetic biological systems with specific functionalities without too many trial-and-error cycles.
Synthetic biology applications often rely on AI simulation models of complex chains of biochemical reactions, which provide advanced insights into how biological systems would behave before they are actually constructed. Through simulations, it becomes possible to represent all biomolecular interactions involved in processes like the transcription of genes when cells reproduce or the translation of genes to proteins. Moreover, when designing a novel biological system that does not exist in nature yet, AI-driven simulations are the only way to understand how those systems would behave.
Biological molecules are so huge and complex that even by knowing their structure, it is not easy to know how they behave. One of the most practical applications of AI is, therefore, predicting the behaviors of segments of proteins, DNA, RNA, or mRNA (messenger RNA, famous because it was used in the COVID-19 vaccines) based on their sequence.
Deep neural networks stand out as the optimal Machine Learning algorithms for synthetic biology, given their scalability and capacity to replicate complex nonlinear relationships between inputs and outputs. In particular, Convolutional Neural Networks (CNNs) are employed to predict how these genomic segments would function. In the previous chapters, we mentioned that Recurrent Neural Networks (RNN) are used for sequence data, and CNNs are used for images, which appears to be a contradiction with this information but is not. CNNs work well for molecules and genome sequences because their functionality depends on their shape; in that sense, molecules resemble an image with a particular shape.
Finally, labeling synthetic biology data so that AI algorithms can be trained is frequently more costly than in other domains because it demands specialized knowledge and, at times, end-to-end data collection processes in the laboratory. This high cost poses a challenge, especially for deep learning models that depend on extensive training data. To address this issue, there is a growing research interest in labeling data automatically and in generating simulated training data using Generative AI. Simulated or synthetic data, though not derived from actual biology experiments, could be realistic enough to train neural networks effectively. The ongoing research on Artificial General Intelligence (AGI) also places substantial emphasis on automatic AI feedback, labeling, and generation of synthetic data. We covered this in detail in Chapter 9.
Sequencing and Synthesizing Genetic Material
This section delves into some of the applications of synthetic biology, starting with sequencing and synthesizing genes.
The precise order of nucleotide bases in a DNA molecule is called DNA Sequencing. Scientists started mapping the human genome — our DNA sequence — in 1990, and after 13 years of work, the whole human genome was successfully mapped by 2003. Once that gigantic task was completed, subsequent initiatives focused on mapping the genomes of other organisms, ranging from fruit flies and mice to economically important crops. The outcomes of large-scale genome sequencing projects contribute extensively to our understanding of genetic diversity, evolution, and the molecular basis of many diseases [Nurk et al.].
Apart from sequencing the DNA, it is also possible to synthesize DNA with a sequence dictated by a template. This is called synthetic genomics. After the synthesis process, the newly created DNA molecules are assembled into complete genomes and transplanted into living cells, effectively substituting the genetic material of the host cell and altering its metabolic processes. This approach first demonstrated its potential by synthesizing multiple virus genomes like hepatitis C and polio in 2000 and 2002, respectively. Notably, among the earliest organisms generated in this manner were infectious viruses [Couzin].
Synthetic genomics has traditionally been very expensive, but recent advancements have enabled cost-effective, large-scale modifications of genetic material. Firstly, Polymerase Chain Reaction (PCR), a laboratory technique used to amplify and produce copies of a specific DNA sequence, has become a fundamental tool in molecular biology. The term PCR has become unfortunately familiar to us all, primarily due to the Covid-19 pandemic in 2020. The second advancement is DNA mismatch error correction, which identifies and repairs errors that occur during DNA replication when a cell divides. Finally, a third advancement that helps reduce cost is the use of modularity engineering principles. Modularity involves the use of standardized components, such as DNA parts, that can be easily extracted, replaced, or combined to create diverse biological functions. This approach enables the construction of complex biological systems by assembling smaller, interchangeable modules, facilitating the design and modification of biological systems [Beardall].
DNA synthesis has advanced significantly, enabling the creation of complex sequences. It is even possible to encode random digital information into synthetic DNA. In 2012, George M. Church codified one of his books into DNA molecules. The book had a staggering 5.3 megabytes of data [Church]. Despite the extravagance of encoding his book, George M. Church is one of the most respected synthetic biologists. The quotation we chose to open this chapter belongs to him.
The molecule encoded by Church was obviously not functional, but AI can naturally be leveraged to design genes that could be fully functional. For example, a deep neural network was created in 2020 to synthesize genomic segments. This AI program was used not only to optimize fitness for function but also to ensure sequence diversity. Sequence diversity is important because it underlies the adaptability and resilience of populations, allowing organisms to evolve, resist diseases, and thrive in changing environments. This AI algorithm ensured sequence diversity through a similarity metric that penalized excessive similarities beyond a set threshold. [Linder et al.].
AlphaFold and The Protein Folding Problem
As hinted above, protein structure prediction is another area where deep neural networks’ intense learning plays a significant role. One of the most celebrated examples is AlphaFold, developed by DeepMind [AlphaFold]. DeepMind is the same company that created Alpha Go, the AI system that defeated Lee Sedol in 2016 in the ancient Chinese game of Go, as discussed in Chapter 7.
Protein design is a complex challenge. These fundamental building blocks of life are sequences of amino acids that naturally fold into distinctive 3-D configurations, which are specific to their biological function. The “protein folding problem” refers to the challenging pursuit of deciphering how the sequence of amino acids in a protein dictates its 3D structure, which entails the spatial arrangement of atoms within the protein molecule.
Understanding a protein’s 3D structure is essential for unraveling its physical function, interactions with other molecules, and potential disease involvement. Protein structure prediction holds immense significance in drug discovery, as it enables the design of pharmaceuticals that target specific proteins involved in specific conditions, offering insights into treatment strategies and developing novel therapeutics.
Before AlphaFold arrived, some experimental protein structure prediction techniques had already been used, like cryo-electron microscopy and X-ray crystallography. However, these methods, while invaluable, were expensive and time-consuming, leading to the identification of only a fraction of the millions of known proteins across various life forms.
In 2018, AlphaFold stunned the scientific community by topping the rankings at the global competition called CASP (Critical Assessment of Techniques for Protein Structure Prediction). It excelled in predicting the most challenging protein structures, even in cases where no templates from similar proteins existed for comparison. AlphaFold participated again in the same competition in 2020, and its performance was even more astonishing. It achieved accuracy levels previously deemed unattainable. AlphaFold demonstrated an unprecedented ability to predict protein structures with remarkable precision, with scores surpassing 90 for two-thirds of the proteins.
The way that it was able to create consistent breakthroughs is necessary to mention. In order to find the 3D structure of a protein, AlphaFold compares it with multiple proteins of known structure that share similarities. Then, it engages in Machine Learning across three diverse data structures of the proteins: a representation at the sequence level, a portrayal of pairwise nucleotide base interactions, and the 3D structure of the protein at the atom level, which the model generates as output.
Despite these groundbreaking accomplishments, protein design remains a multifaceted puzzle. AlphaFold success is a giant leap forward, but challenges persist. For example, AlphaFold primarily focuses on single-chain proteins, leaving out extremely complex or intrinsically disordered proteins. Additionally, questions linger about how well AlphaFold can predict totally novel folds or structures that are utterly absent from existing databases.
Engineering Proteins with Generative AI
What AlphaFold does is find the shape — and, therefore, the function — of a protein sequence. Doing exactly the opposite is also an important application of synthetic biology, namely finding the sequence of a protein for a specific function, which basically means designing new proteins.
Designing proteins is a formidable task. Expert Systems with simple Machine Learning and rule-based algorithms have been traditionally used, but progress is self-limiting, and require specialized knowledge to identify the features that contribute most to performance. Recognizing the limitations, there is a lot of current research on Generative AI algorithms such as Generative Adversarial Networks (GAN) and Variational Autoencoders (VAEs), which we covered in Chapter 5. Using these AI models, researchers can work backward from the desired biological functions of the protein and to the DNA sequences that produce a protein that fulfills these functions [Tucs et al.].
Synthetic biology allows creating new protein configurations that match or even surpass the existing proteins in terms of specific functionalities. For example, Hemoglobin, the protein responsible for transporting oxygen in blood cells, can also bind to carbon monoxide, leading to the health risk of carbon monoxide poisoning. In 2009, a research group created a helix bundle mimicking hemoglobin’s oxygen-binding properties but without this affinity for carbon monoxide, substantially reducing the risk of poisoning [Koder and Anderson].
Moreover, by manipulating protein structures and functions, it is possible to design industrial enzymes for many applications, like improved detergents and lactose-free dairy products. Synthetic biology also offers promising avenues in biochemical production for various industrial uses. Biological cells can be used as microscopic molecular factories to generate materials, typically proteins, with genetically encoded properties.
One example is the production of proteins required to grow biofilms. A biofilm is a slimy, adherent layer of microorganisms that forms on a surface, often in water, and is characterized by a protective matrix that allows bacteria to adhere and thrive. Biofilms have diverse applications. In healthcare, biofilms are cultivated on medical devices to either protect against infection or, conversely, combat bacterial contamination. In industrial contexts, biofilms contribute to processes like wastewater treatment and bioenergy production. Additionally, biofilm communities influence nutrient cycling and pollutant degradation in natural ecosystems.
Synthetic biology, when it adopts AI, expands and speeds breakthroughs, including mapping existing proteins and creating new, non-naturally occurring proteins. Companies are actively deploying AI-informed models and processes to productize their learnings and IP about proteins, paving the way for organ development, human tissue and function enhancements, and gain-of-function.
The Future of Transplants with 3D Bioprinting
Putting many proteins together and building an organoid with them is another of the applications of synthetic biology.
Organoids are artificially cultivated organs like an artificially grown kidney or liver. In the previous chapter, we talked about using 3D printers to print mechanical organs like bionic hearts. In this chapter, we talk about 3D bioprinting instead. Organoids are typically printed in a laboratory setting using this kind of printing technology. 3D bioprinting is a process through which cells are layered or deposited in a specific pattern to create three-dimensional structures that mimic the architecture of natural tissues or organs. These bioprinted organoids can serve as models for studying biological processes, drug testing, and potentially for transplantation into a patient [Hong].
Before they are bioprinted, living cells need to be created. These cells are artificial cells created from lipid vesicles through synthetic biology procedures, and they encompass all the essential components required for a functioning cellular system. These artificial cells are created with all the functional and design characteristics necessary to be considered alive, including self-replication and self-maintenance.
The next stage of the process is bioprinting these artificially created cells into an organoid, a very delicate process. Bioprinted organoids must meet intricate shapes and complex requirements. For instance, a bioprinted heart must meet structural criteria such as mechanical load, vascularization, and electrical signal propagation requirements, among many others.
Although 3D bioprinting is a very new technology, it has already demonstrated its effectiveness. The first successful transplant of a 3D bioprinted organ created from a patient’s own cells was reported in 2022. This pioneering procedure focused on reconstructing an external ear to treat microtia, a congenital condition characterized by an underdeveloped or malformed outer ear [Clinicaltrials].
The implications of bioprinting for humankind are immense. In the near future, bioprinting will offer a range of possibilities, including the production of bioprinted eyeball corneas and hair follicles, structures less complex than, say, a bioprinted heart but more delicate than a bioprinted external ear. In healthcare, bioprinting will allow for the fabrication of organs in organ farms, addressing the challenges posed by organ shortages for transplantation. Patients’ own cells will be employed to craft miniaturized organs or organoids for testing medical responses before administering treatments. Bioprinted prosthetic limbs could be meticulously designed to seamlessly integrate with the individual’s body, enhancing their functionality and comfort.
Finally, bioprinting is also paving the way for the implantation of electronics into biological tissues, resulting in cybernetic organs that surpass their natural counterparts in sophistication. For instance, bio-printed lungs could be equipped with sensors and nano-filters that purify the air before it enters the bloodstream. Similarly, replacement bionic eyeballs could come with built-in zoom and infrared vision capabilities.
The implications of printing electronics into biological tissues that would later be integrated into a human are breathtaking. This kind of technology could surpass the possibilities of the BCIs presented in the previous chapter and bring about a truly interlaced living organism with a biology designed by AI. We see this as solving, at least directionally, core mechanical issues around interlacing, the nuts-and-bolts of how it can be physically implemented.
Creating Synthetic Life and Non-Natural Life
3D bioprinting of organoids is impressive, but organoids themselves are not living per se despite having the necessary design and construction properties to claim otherwise. One polemic facet of synthetic biology is the overall concept of synthetic life.
To set a definition, synthetic life involves the construction of organisms in a controlled setting using synthesized molecules. Experiments in this field serve various purposes, including exploring the origins of life, testing innovative therapeutic and diagnostic solutions, investigating fundamental life properties, and finally, the ambitious endeavor of generating life from non-living components [Deamer].
This artificial form of life is called Synthetic Life, a term coined by Craig Venter in 2010 when he created a fully synthetic bacterial chromosome and introduced it into genetically depleted bacterial host cells. Four “watermarks” were incorporated into the DNA of the single-celled organism to help identify it: a complete alphabet code table with punctuations, 46 contributing scientists’ names, three sentences, and the secret email address for the cell. Remarkably, these synthetic bacteria demonstrated growth and replication capabilities. [Gibson et al. #]. Since then, increasingly advanced synthetic life organisms have been engineered to perform essential functions, such as pharmaceutical production or detoxifying polluted land and water sources.
Going beyond the creation of synthetic life, synthetic biology also ventures into the realm of unconventional molecular biology, which is creating life that could not possibly exist in nature. In nature, across all living organisms, there are only five nucleotide bases and 20 amino acids, but AI makes it possible to target specific chemical and biological properties to design and engineer non-natural proteins and non-natural bases that cannot be found in nature. This would solve the “first building block” issue of designing non-natural forms of synthetic life that use these novel amino acids and bases that do not exist in nature.
While such non-natural organisms might offer advantages, they also pose unique risks. Once released into an environment with ecosystems that have traversed development for millions of years, these organisms might be able to engage in horizontal gene transfer or gene exchange with natural species, leading to unpredictable initial outcomes.
It is also not clear if these species could nourish themselves on natural proteins or if they would have to rely on non-natural molecules. If these new species were engineered to depend on non-natural materials for synthesizing their own proteins or nucleic acids, they would be incapable of surviving in natural environments if inadvertently released. Conversely, if they manage to feed on natural molecules and establish themselves in uncontrolled environments, they could potentially outperform natural organisms, resisting predators and biological viruses, leading to uncontrolled proliferation, which could include extinction for some current species.
Any discussion of synthetic life created from molecules that do not exist in nature might seem like Science Fiction, but it is not. The first living organism with this kind of non-natural DNA was unveiled in 2014. Researchers added two new, unconventional nucleotides to bacterial DNA, and the newly created non-natural bacteria underwent 24 generations of growth, all containing the newly introduced artificial nucleotide bases [Malyshev et al.].
Xenobots: from Frog Embryos to Synthetic Lifeforms
An example of synthetic life is the Xenobots. Xenobots derive their name from the African clawed frog Xenopus Laevis and were developed in 2020 [Sokol] [Simon]. Xenobots use the conventional molecules found in nature, not the non-natural variations we just talked about.
Xenobots are typically less than 0.04 inches in width and comprise two primary parts: skin cells and heart muscle cells. These two kinds of cells are obtained from embryonic stem cells from these African frogs. The skin cells give structural support, while the heart cells function as miniature engines that contract and expand rhythmically, pushing the xenobot forward. The specific arrangement of a xenobot’s body is determined through computer modeling, employing a trial-and-error process.
Whether xenobots should be classified as living organisms, robots, or something completely different is a matter of ongoing debate among scientists. However, it is clearly demonstrated that xenobots exhibit some common characteristics with living organisms. First, xenobots can self-repair when injured. Second, xenobots can reproduce by gathering free-floating cells from their surroundings and assembling them into new xenobots with the same capabilities. Finally, xenobots can survive for extended periods without nourishment.
Xenobots have been engineered to perform various tasks, including walking, swimming, pushing pellets, transporting payloads, and collaborating in swarms to gather scattered debris into organized piles on their dish’s surface. Based on these behaviors, xenobots hold promise for some practical future applications. For example, they could gather ocean microplastics into larger masses for easy removal and transport to recycling facilities. Additionally, in clinical settings, xenobots may be employed to address tasks such as removing arterial plaque and treating medical conditions.
Biological Computers and Biological Machine Learning
In case the creation of synthetic life — either natural or non-natural — proves insufficiently ambitious, contemplate the prospect of developing biological computers. A biological computer is a specially designed biological system with the ability to perform operations similar to those in electronic computers. Scientists have already developed and characterized various logic gates in multiple organisms like their electronic counterparts in computers.
Recent advances show that it is possible to integrate analog and digital computation circuits into living cells. For example, in 2007, a biological circuit that could execute logical functions like AND, OR, and XOR was implemented in mammalian cells [Rinaudo]. Moreover, in 2011, scientists developed a therapeutic strategy employing biological digital computations for detecting and eliminating human cancer cells [Xie]. In addition, in 2016, computer engineering principles were applied to automate the design of this kind of digital circuitry within bacterial cells [Nielsen], and in 2017, arithmetic and Boolean logic were also implemented in mammal cells [Weinberg].
On top of building these more simple computational circuits, it is also possible to implement complex artificial neural network analogs using biomolecular components. The implications of this include enabling the execution of intricate machine-learning processes within a biological system. Furthermore, in 2019, a theoretical architecture for a biomolecular neural network was presented [Pandi et al.]. This architecture is a network of chemical reactions that accurately executes neural network computations and demonstrates its use for solving classification problems. It serves as the biomolecular equivalent to the perceptron built in 1957 by Frank Rosenblatt, with a similar simple structure composed of one layer of artificial neurons.
The implications of being able to build machine-learning algorithms inside biological systems are enormous because that means that it would be possible to increase the intelligence level of living beings, even humans, by following computer architectures constructed with biological materials.
Masters of Creation
Synthetic biology is arguably the most revolutionary application of AI. It has the potential to transform the essence of human beings by altering their biological fundamentals, allowing them to live longer, resist diseases, and acquire superhuman abilities with specifically designed organoids. Driven by the power of AI, it will occur alongside the development of entirely new biological life forms. Advances such as integrating electronic or machine-learning elements into biological circuits solve thorny mechanical issues regarding how interlacing will occur and open a solid avenue for human-AI interlacing.
The huge power of synthetic biology makes us the masters of creation. Or maybe the masters of disequilibrium. We will discuss this in subsequent chapters.
Buy “Machines of Tomorrow” on Amazon
Author: Pedro Uria-Recio
Join our substack
Check us out at machinesoftomorrow.ai
