AI in an Era of Synthetic Life

By Michael Megarit 

Proteins, the building blocks of life, make up 75% of your dry body weight. Proteins play an essential role in our physical being and carry out numerous vital tasks from muscle formation and blood clotting to hormone secretion and hair production – understanding them is like unlocking one major aspect of biological understanding and control.

Still, one major challenge remains. Deciphering DNA sequence does not reveal the function of proteins; understanding their folding process is the key to unlocking their full functionality. Each protein structure, formed through complex folding, plays an essential part in its purposeful functioning. Collagen fibers in the tendon resemble rope, while enzymes are designed to adapt their structures according to which molecules they interact with. Predicting protein folding was once an impossible feat, with traditional computation methods often failing to reach their full potential in their pursuit of possible shapes of proteins. This prevented advancements across many fields ranging from pharmaceuticals to enzymes capable of degrading plastics from taking place.

As part of their search for solutions, in 1993 the Critical Assessment for Structure Prediction (CASP) competition was initiated, in the hopes of solving protein folding puzzles. This event became an indispensable source of competition among experts within this highly specialized yet tight-knit domain; nevertheless despite steady progress no breakthrough had yet occurred.

At CASP13 in Cancun in 2018, DeepMind’s AlphaFold, which began life from a week-long hackathon in 2016, outshone 98 experienced teams. This marked an extraordinary breakthrough for computational biology as AlphaFold employed deep generative neural networks to predict protein folding from their DNA sequences – outperforming competitors by accurately predicting 25 of 43 complex protein structures within mere hours! This stunning victory illustrated just how quickly AI and biotech have advanced.

AI-driven methods like this one, which did not rely on conventional pharmaceutical expertise or computational techniques, marked an impressive accomplishment of AI in biology. Renowned experts such as Mohammed AlQuraishi were left marveling over this feat; marking a new era where AI is not just used as an aid but rather plays a key role in unlocking biological secrets.

Two years later, their team came back with another groundbreaking development that made headlines worldwide. Scientific American captured it best: “One of the Biggest Problems in Biology Has Finally Been Solved.” AlphaFold had revealed previously obscure proteins with incredible speed – rendering CASP competition obsolete in much the same manner as ImageNet had done so long before – thus solving a challenge which had baffled scientific communities for over fifty years!

AlphaFold2 for public use was undoubtedly a game-changer in 2022, catalyzing an explosion of advanced machine learning applications in biological research – both fundamental and applied. Dubbed by one researcher as an “earthquake,” over a million researchers worldwide used AlphaFold2, including top biology labs globally – within 18 months it had been accessed by millions – from antibiotic resistance studies, rare disease treatments, to origins of life exploration. Prior to its public launch only approximately 190K protein structures had been identified and recorded within Europe Bioinformatics Institute database; after its upload by DeepMind (around 200 Million structures added) nearly covering almost the entirety of known proteins!

Determining the shape and function of proteins was once a laborious, time-consuming task; now it begins in seconds – marking exponential change and opening up an endless stream of possibilities.

This breakthrough marks just the start of an unprecedented integration between two revolutionary technologies. Bio-revolution and AI advances are coevolving, with future innovations dependent on both technologies. Imagine two great waves combining into one massive wave: this could be our future! Artificial intelligence and synthetic biology may appear similar; after all, all intelligence known today comes from life itself – though different fields focus on different aspects of this reengineering project that leads to human advancement.

Due to the vast amounts of biological data such as protein structures, biological researchers need new tools that go beyond traditional approaches. Teams are currently creating products which generate new DNA sequences using natural language instructions. Transformer models are helping us untangle the complexity of biology and chemistry, unlocking relationships and meaning in sequences previously unintelligible to human minds. Language models calibrated with biochemical data are offering plausible new molecular structures, protein sequences, DNA and RNA sequences that have yet to be validated in laboratories, and predicting various properties of compounds which later get validated in laboratories. Applications and research for such models is increasing at an accelerating pace.

Explorations into merging human consciousness with digital systems are making progress. A landmark 2019 procedure saw electrodes implanted into the brain of an advanced ALS patient to allow him to convey his heartfelt message of love for his son: “I love my cool son”. Companies such as Neuralink are creating technologies for brain-machine interfaces in order to establish direct connections between humans and computers – testing ultrathin hair-like electrodes on pig brains in order to monitor neural activity before eventually starting human trials with its N1 brain implant shortly. Meanwhile Synchron has initiated similar trials in Australia – both have begun human trials as well.

Cortical Labs has taken another significant step by cultivating neurons in vitro and building an artificial brain capable of playing Pong. Carbon nanotube neural “laces,” designed specifically for this task, may one day link our minds and digital world together seamlessly.

Human minds having immediate access to the vast computational power and data resources provided by the internet and cloud computing is inconceivable at first, yet researchers are making great strides forward. Artificial Intelligence and synthetic biology, two pivotal technologies of the coming era, are inextricably linked with one another in an ever-accelerating feedback loop – each spurring progress of its counterpart. While pandemic has raised visibility of biotechnology advancement, its implications – along with potential and challenges associated with synthetic biology research are only now beginning to seep through into public awareness.

At the dawn of a new epoch, biomachines and biocomputers will become the norm, with DNA performing computations and synthetic cells harnessed for tasks. Welcome to an age of synthetic life where machines embody life.