The framework conceptualizes GRNs as agents actively navigating the transcriptional space and provides a set of tools, leveraging computational models of curiosity-driven learning and exploration, with a battery of empirical tests inspired from behaviorist tradition, for automated experimentation and behavioral characterization. The proposed framework is novel in two central ways. First, it introduces a novel AI-based toolbox to the field of biological network analysis. We show how this toolbox, leveraging the successful ingredients of recent intrinsically motivated learning algorithms – originally developed to enable robotic AI agents to explore and learn diverse skills in novel and unstructured environments (Baranes and Oudeyer, 2013; Forestier et al., 2022) – can be transposed to assist efficient discovery of behavioral abilities within biological pathway models like GRNs. Second, rather than merely mapping the attractor states (Kauffman, 1993; Kauffman, 1995; Dang et al., 2011) or analyzing their sensitivity to model parameter changes (Ingalls, 2004; Ingalls, 2008) as extensively proposed in conventional GRN analysis methods, our framework investigates the dynamic adaptability of these networks’ navigation competencies in response to various changing environmental conditions. – AI-driven automated discovery tools reveal diverse behavioral competencies of biological networks; Computational and Systems Biology; Jan 13, 2025; Etcheverry, Moulin-Frier, Oudeye, and Levin.

Steven Cole stood in front of the crowded conference room smiling at the visiting audit team from the Inspector General’s office. The morning presentations went well, and he noted with smug satisfaction that most of the visitors only feigned understanding of the project, some visibly bored. The one exception was a young woman in a dark suit sitting in the first row, taking notes, attending to his every comment. Her direct stare unsettled him.

Each visitor carried a thick binder labelled “High Mobility Extreme Combat Infantry”, all marked “Top Secret”, with serial numbers and counter-signatures. Bruno Abrams stood in the back, hands clasped in front of him.

“I hope everyone found the tour informative. Before we take you to the training facility, I thought we should answer any questions you may have from the briefing binders. Perhaps it might be useful for you to hear some background on the theoretical advances that are the foundation of what we’ve achieved here. The Biological Combat Systems Program has developed new techniques that, when Dr. Abrams and I are permitted to publish, will dramatically advance our understanding of human genomics.”

He paused and looked around the room, reading the crowd. He suppressed a flash of irritation at the lack of response, then continued.

“First, let me recapitulate the ‘why’ of the BCS program before we expand on the ‘how’, which Dr. Abrams will cover. With the latest international treaties, the ongoing Military Technical Revolution in battle networks has reached a plateau. The rapid increases in the precision, lethality, and autonomy of machine systems continuously shifts the balance on the battlefield, provoking countermeasures that seek either to overwhelm with mass, or compromise command and control by cyber measures. All of that pits machine against machine, in an escalation loop that humans risk loss of human control. The recent pause opens up a window to possible solutions that require highly mobile infantry capable of operating in extreme combat environments. Hence, the Biological Combat Systems program.”

Cole interpreted the silence in response as permission to continue.

“Working alongside the Autonomous Combat Systems programs, we explore ways to develop and deploy new capabilities that cannot be hacked or overwhelmed, to complement and augment human infantry, while observing current treaty constraints on human experimentation. Our HiMEC project has done that, all while complying with existing restrictions governing genetic engineering. Because we know our adversaries are not abiding by many of the treaty restrictions, we believe these capabilities are essential to maintain battlefield superiority. Dr. Abrams will give you an overview of our work, we’ll take questions, and then we’ll conclude the visit with the training facility. Dr. Abrams?”

Bruno Abrams stepped forward and gestured at the screen, calling up his presentation. An image of a strand of regulatory RNA, curled back on itself in twists and folds appeared.

“Biology is profoundly non-linear. A specific organism constitutes a solution space, and what we do with our work is boundary shaping.” As a group, the auditors leaned forward. Abrams continued. Cole frowned.

“When the human genome was first completely sequenced decades ago, we believed we were close to a total understanding of human biology. How wrong we were. Knowing only the identity and sequence of the nucleotides in one person’s genome is akin to knowing the text of War and Peace, in order, but without punctuation, spacing, or formatting. But that’s not enough. You don’t know the meanings of words. The rules of grammar are missing. You don’t know the cultural context. You don’t have the background knowledge of human nature and relationships that give the novel meaning. And you don’t know what variations might have been introduced in various translations from Russian to English. That’s the challenge we faced with the genome back then. We needed to develop the vocabulary, the grammar, learn the language of the genome, and quantify the enormous amount of individual variation across the human population. As that work progressed, we then had to place the story told by an individual genome into the context of the species, the local environment, the specific history of that organism, and its relationship to other organisms in its local genosphere.”

The woman in the front row raised her hand. Abrams nodded to her.

“Ms. Jackson?”

“How are you using the term genosphere in this context?”

Abrams smiled.

“Excellent question. The genosphere is the association of individual genomes and how they interact and influence each other. The term originated in the ‘90’s with Russian biologists, and really caught on when the field of epigenetics exploded and demonstrated the deep connections between genomic activity, regulation, and environmental influences.”

“Are you referring to Vernadsky? He was actually Ukrainian,” she confidently asserted. Abrams smiled at her obvious enthusiasm and familiarity with the topic.

“Actually, I was thinking of Sauchanka, who was also Ukrainian, but more specifically the bench researchers who expanded on those concepts. But I’m impressed. Thank you for that clarification.” The woman looked down, embarrassed by the attention.

Abrams continued. “It’s the context. The full meaning, the expression of the information, in the genetic sequence can’t be fully appreciated or realized unless it is in the appropriate context. Knowing how to manipulate an individual genome, or even a specific gene, requires knowing the context in which that genome operates, it’s history, influences, relationships. It was the mathematical formalization of those insights combined with the use of AI tools that allowed us to take the next step.”

The young woman raised her hand again.
“When you say context, do you mean where the information is read, or when?”

Abrams smiled, apparently enjoying her inquisitive enthusiasm. Cole frowned.

“Both, as well as what other associated systems are interacting. There are many layers involved.” He gestured at the screen, and the twisted molecule was joined by another image, of a strand of DNA.

“At its most basic level, the regulation of gene expression is an exercise in molecular geometry, the mathematics of the complex surfaces and shapes created by the quantum interactions of electrons. Look at this piece of RNA, which regulates the expression of a particular gene, and how it combines with the target site.” The image moved, with specific surfaces on each molecule lighting up as the figures rotated and approached each other, the lighted areas mating like three-dimensional puzzle pieces.

“Binding at this site blocks the expression of a particular gene. Other binding sites promote the expression of other genes. But it all comes down to two or more molecules fitting together in a certain way, like two puzzle pieces or more appropriately, a key into a lock.”

A few of the auditors nodded, so Abrams plowed ahead.

“So what determines the shapes of these molecules? Each one is made up of a variety of basic building blocks called nucleotides, or amino acids in the case of proteins, each of which has a particular shape. In various combinations, they can twist back on each other and fold into more complex shapes. All of those interactions are driven by the contours of the electron clouds in each atom, the thermodynamic relationships between the atoms, external factors like the acidity of the local environment, external electrical charges, that sort of thing. This is where the key insight occurs.

“Each of those variables constitutes a dimension of a complex mathematical space and using enhanced AI tools we’ve developed techniques to map the physical shapes of the larger molecules onto the various physical parameters of their constituent parts. This enables the AI to test candidate molecules, but more importantly, it gives us a theoretical basis for understanding higher regulatory processes.

“The simplest understanding of how the information in DNA is transformed into a living thing is that it is transcribed into RNA, then that RNA is translated into proteins, and these proteins are the basic constituents of living things.

“The complexity comes in with how the DNA is transcribed. At each step there are Gene Regulatory Networks, or GRNs, that create feedback loops that both inhibit and accelerate various processes, local feedback, mid-level feedback, and external feedback. For example, the RNA and protein products of gene expression both can create regulatory molecules that interact with DNA to both enhance and suppress gene expression. Those molecules can be further edited to create additional layers of regulatory complexity. Gene regulation is where the vast majority of the diversity of living organisms comes from, not necessarily the individual genes themselves.”

Cole saw confused looks on some of the auditors and interrupted.

“This is why we have had such great success with the animal projects, working our way up from smaller mammals up to primates. As you’ll see, we’ve been able to greatly amplify the capabilities of our primate platform, while completely avoiding the regulatory and ethical issues surrounding human genetic engineering.”

The lead auditor was shaking his head.

“I’m sorry, you lost me a few steps back.”

Cole continued, stepping in front of Abrams.

“Let me give you an analogy. Take a musical score, a sequence of notes on a piece of paper that communicates a certain melody. That melody can be played fast or slow, with a full orchestra or with a jazz band, with various accompaniments, harmonies, and in the hands of different conductors, producing musical experiences that are vastly different. The genome is a musical score, and it’s the regulatory processes that produce different organisms.”

“And the musical score also contains the instructions for building the musical instruments,” the inquisitive young woman interjected. Abrams nodded enthusiastically.

“Precisely. In addition…”

Cole jumped in again.

“The difference between the gorilla genome and humans is actually quite small, and that’s where we’ve achieved our greatest success. By manipulating the regulatory pathways, we’ve produced incredible results.”

Abrams shook his head impatiently, visibly bristling at the interruptions.

“Each step in gene expression has regulatory feedback loops at multiple levels,” Abrams continued, “locally inside the nucleus, within the cell associated with other intracellular processes, then more generally inside the organism, then finally from the external environment. It was understanding and formalizing the interplay between those different levels of genomic regulation that allowed us to make our most important advances.”

He paused, waiting for his words to sink in, looking from one person to the next. Cole fidgeted next to him. A woman in the back raised her hand.

“How is this different from what the Chinese and the British do?”

Cole stepped forward waving a finger.

“Yes, let’s address that. First, our techniques are far more sophisticated, and that’s not just boasting. The technology up until now has been comparable to using sledge hammers to fix a computer. Our methods are far more subtle and precise, allowing us an unprecedented level of control. But what you are really asking about, if I may, is the ethical concerns, correct?”

The woman dipped her head. Cole continued.

“The British Human 2.0 catastrophe held lessons for all of us in the amplification research community. But even the Chinese manipulations of domestic pets have raised sufficient concerns that we are watching them closely. Of course, no one is fooled by the purported entertainment value of their work. The PLA research facilities are deeply involved in the Chinese efforts, and we understand from our military colleagues that the Chinese are aggressively pursuing military applications. Hence, the interest in our work by our own military. We are supervised very closely.”

Finally the lead auditor spoke.

“This is quite a bit to digest and is very impressive. No doubt your team has achieved some remarkable things, but that’s not why we’re here, or at least not the only reason. In addition to auditing the expenditures, we’ll also be evaluating your results, particularly whether you are complying with bioethical regulations. This is new territory, gentlemen, and I, for one, am not comfortable with the oversight the…” he glanced down at his binder, “…High Mobility Extreme Combat Infantry program has received up until now.”

Abrams flushed, but made no reply. Cole jumped in once more.

“Dr. Abrams and I are very proud of our accomplishments with HiMEC, and as you will see, we have gone well beyond existing regulations to ensure we are fully compliant. I don’t think I need to remind you of Dr. Abrams’ credentials as a former member of the National Academy of Sciences bioethics panel and his early breakthroughs with somatic stem cells. That changed the debate around stem cell research and enabled many important medical breakthroughs.” Cole paused and put a hand on Abrams’ shoulder. “If Dr. Abrams’ work is good enough for the Vatican, I’m certain the United States government will find it worthy.” Cole smiled around the room, as Abrams hung his head, jaw clenching, all but shrugging off the weight of Cole’s hand.

Another auditor raised her hand.

“Regarding results, I’ve spent some time reviewing the testing data you’ve included here, and beyond the general achievement of amplifying gorilla capabilities, I don’t see anything notable from an operational perspective. The intelligence testing indicates the subjects are generally of average human intelligence. Of course, that is a remarkable improvement for a non-human primate, but how does that translate into an enhanced battlefield capability?”

Cole’s smile tightened.

“We’ll address some of those questions at the training facility. In short, the accelerated nerve conduction, enhanced muscle density, and particularly the cutaneous chromophore enhancements create multiple advantages in a tactical environment that we are in the process of elaborating.”

Another auditor raised his hand.

“John Sutter from the Hastings Center. I’m interested in hearing more about the religious behaviors alluded to in the section about their language skills. Is that part of the training plan, or is it something else? I’m not sure how that enhances any operational capability, either.”

Before Cole could respond, Abrams stepped forward quickly and interrupted with a raised hand.

“That was an unexpected development. Their facility with languages created the need for materials in other languages, from other cultures. The selections included religious texts as well as collections of myths and legends that were so enthusiastically received, they enabled accelerated training by using them as rewards and inducements, but this led to the unintended consequence of eliciting spiritual explorations.”

Cole scowled and took back control of the conversation.

“The subject’s religious syncretism hasn’t really impacted their military training, and has really become more of an entertainment, a distraction they pursue during down time.”

Sutter nodded, but appeared unconvinced.

“I’d like to learn more about that.”

***

Several assistants ushered the group out of the conference room, and Abrams excused himself. Cole accompanied the tour group. The team quickly moved through the laboratories and computer facilities, eager to see the test subjects. Finally, they gathered around a large bank of thick angled plate glass looking out over a courtyard filled with several dozen milling figures, in groups wearing colored loincloths denoting teams.

The auditor’s faces went pale, eyes wide. Several walked up to the glass and leaned forward. One man became unsteady and had to sit down in a chair, bowed forward, head in his hands. A petite woman sat with him, rubbing his back, looking over at the observation window anxiously. The lab technician leading the tour looked to Cole for guidance, who made a small hand gesture for him to continue. Cole sat next to the woman and whispered explanations to her while the man collected himself.

The technician continued his presentation.

“This is one of our training areas. You can see they mostly work in small groups, and we now rely on the older subjects to mentor the younger ones. Frankly, the human instructors are no longer capable of keeping up with them, and are now mostly utilized for monitoring and supervision, security, actually.”

He pointed to the equipment around the courtyard. High concrete walls enclosed the space, with cameras affixed to the walls at intervals. A tower loomed at the far corner, tinted windows concealing the interior. On top, the two .50 caliber machine guns at the corners were discretely covered.

In the yard, some of the teams ran through exercises, practicing drills and climbing on equipment. Some gathered around a speaker leading instruction. Several sat off to the side hunched over tablets, engrossed in reading save for the occasional tap on the screens to change a page or look up an unfamiliar expression.

“What are those ones doing?” the IG representative asked, pointing to the readers. The tour leader spoke up.

“Those are the reinforced reading tablets we mentioned before. Reading time is a perk, a reward for successful task completion or good behavior. The tablets are occasionally a source of conflict, which is why we had to make a significant investment in enhancing the units to handle the rough treatment. Their talons put a real beating on the touch surface.” The technician grinned, without any response from the delegation.

Cole scowled at him and the expression disappeared.

“What are they reading?” Sheehan asked.

“The current fad is Shakespeare. Dr. Abrams authorized us to upload the collected works a few months ago, and they are devouring them. We’ve also added a few videos of plays, and they love them. It’s a very powerful behavior modification tool.”

Dr. Cole stepped forward.

“One of the unexpected benefits of the cognitive augmentation is this exceptional language facility. They are very adept at learning languages, and they are extremely fond of stories, storytelling, anything with a narrative. We discovered early on that training objectives are best communicated, and most successfully learned, when they are incorporated into stories. Dr. Abrams first used Greek literature and Norse myths to develop this on both counts, and effectively motivated our earliest successful subjects, the Charlie series, by using those stories. The Iliad was our first major breakthrough demonstrating their linguistic skills and simultaneously teaching martial values of military culture, like honor, duty, and obedience. We’ve continued the practice.”

“How many versions are there? Of the, uh, subjects?”

“The smallest ones you see out there are the Epsilons, the fifth generation. Dr. Abrams has a sixth generation underway as well. Each one contains further refinements, amplifications of some abilities, repression of other characteristics, all of which is translating into spectacular results in field trials. We continue to discover unexpected byproducts of various amplification efforts. For example, by increasing muscle mass manipulating the myostatin regulatory complexes, we inadvertently activated pathways resulting in the claws. The improved nerve conduction using the squid axon sodium channel enhancement allowed for the introduction of the chromophores that create their camouflage.”

“How is that you’ve been able to compress the development time? The older ones out there look like fully grown adults, but they must be only a few years younger than the program itself, correct?”

Cole nodded.

“One of the amplifications is accelerating development. It is risky, though. One reason human development takes so long is the need for gradual myelinization of long fiber tracts inside the brain and around the body. It’s why human infants are so vulnerable. Fortunately, we were able to build on the relatively rapid development of the gorilla nervous system and amplify some abilities to near human levels. They are somewhat deficient in some cognitive domains, mathematics for example, but we’re finding those don’t interfere with tactical performance and mission priorities. We also encountered some issues in earlier versions with accelerated senescence as well, but that appears to have been addressed in current versions.”

The distraction of discussing the training diverted attention from the exercise yard and the group soon moved on to the next phase of the tour.

Chapter 41