Our Vision

Epidecode Amorim-Estrada is a speculative conceptual framework centred on one fundamental problem in modern epigenetics: biological regulation is dynamic, yet most current technologies still measure it through isolated laboratory snapshots.

At present, epigenetic processes such as DNA methylation, chromatin accessibility, histone modification and transcriptional regulation are usually investigated through tissue extraction followed by laboratory-based analysis. These methods include bisulphite conversion, methylation arrays, chromatin immunoprecipitation, sequencing technologies and related molecular techniques. While scientifically powerful, they generally provide retrospective measurements of biological states at a specific moment in time rather than continuous observation of regulatory dynamics as they unfold.

The concept behind Epidecode Amorim-Estrada concerns the theoretical development of an integrated biosensing and computational platform capable of repeatedly detecting selected epigenetic and chromatin-associated regulatory markers longitudinally and, eventually, in near real time..

The proposed framework would not attempt whole-genome epigenetic analysis. Instead, it would focus on a highly targeted panel of biologically meaningful CpG regulatory sites.

CpG sites are genomic regions where a cytosine nucleotide is followed by a guanine nucleotide linked through a phosphate bond. These sites are of particular interest because methylation at CpG regions plays a major role in regulating gene expression. Some CpG regions act as regulatory switches influencing whether genes become more transcriptionally active or more transcriptionally repressed.

Importantly, not all CpG sites are biologically meaningful. The proposed system would therefore prioritise CpG regions associated with known developmental, neurological, immune, metabolic or ageing-related regulatory pathways. Examples could include regulatory regions linked to stress responsivity, neurodevelopment, inflammatory signalling, circadian regulation, cellular ageing, synaptic plasticity, or disease-associated transcriptional pathways.

Beyond DNA methylation alone, the framework also considers chromatin organisation, histone-associated regulation and three-dimensional genomic architecture as interacting components of dynamic biological regulation. Gene expression is shaped not only by isolated molecular modifications, but also by the spatial organisation of DNA around histone proteins, chromatin accessibility and long-range regulatory interactions occurring within the nucleus. Regulatory regions may become more transcriptionally accessible or more transcriptionally constrained depending on how chromatin structure changes across time, developmental states and environmental conditions. Future computational approaches may therefore require integrative models capable of representing multidimensional epigenetic interactions longitudinally rather than through isolated endpoint measurements alone.

The rationale is that dynamic monitoring of a small but biologically informative panel may ultimately prove more computationally and technologically feasible than attempting large-scale continuous genome-wide epigenetic surveillance.

Conceptually, the biological detection system could involve several converging components:

• Microfluidic sampling systems capable of handling extremely small biological samples repeatedly over time.

• Biochemical discrimination mechanisms capable of distinguishing methylated from unmethylated DNA regions.

• Nanosensor or electrochemical detection technologies capable of converting molecular interactions into measurable digital signals.

• Semiconductor integration enabling repeated signal acquisition and miniaturised processing.

• Longitudinal calibration systems capable of distinguishing genuine regulatory shifts from biological noise, sampling artefacts or transient fluctuation.

However, the greatest challenge may not lie solely in molecular detection itself, but rather in the computational architecture required to interpret epigenetic regulation dynamically.

Epigenetic systems are highly multidimensional and temporally unstable. Biological signals fluctuate continuously across development, circadian rhythms, immune activation, environmental exposure, ageing and cellular adaptation. A meaningful system would therefore require software capable not merely of storing molecular measurements, but of modelling temporal biological trajectories.

The envisioned software architecture would involve several layers:

• Signal acquisition and preprocessing pipelines designed to normalise biological input and reduce artefactual variability.

• Pattern recognition systems capable of identifying recurring regulatory signatures across time.

• AI-assisted modelling capable of distinguishing stable biological trajectories from transient fluctuation.

• Longitudinal predictive frameworks capable of identifying whether certain regulatory patterns remain stable, diverge, intensify or regress over developmental periods.

• Dynamic visualisation interfaces capable of representing temporal epigenetic change rather than static molecular states.

• Regulatory-network modelling systems potentially capable of inferring interactions between multiple epigenetic pathways simultaneously.

In this sense, the project becomes as computational as it is biological.

This is also where Alan Turing’s broader intellectual legacy becomes conceptually relevant. Turing demonstrated that machines could formalise and process patterns exceeding ordinary human computational capacity. Epigenetic regulation may ultimately represent a biological system whose complexity requires machine-assisted interpretation not simply because of data quantity, but because the underlying regulatory architecture itself is fundamentally dynamic.

The broader conceptual objective is therefore not merely the detection of isolated methylation markers, but the possibility of constructing machine-readable models of biological regulation unfolding through time.

If technologies of this nature were ever achievable, they could potentially open new ways of studying neurodevelopment, immune adaptation, disease progression, treatment response, ageing, environmental interaction and biological resilience.

At present, however, this remains a speculative and exploratory concept rather than an existing technological platform. Significant scientific, engineering, computational, biological and ethical barriers remain unresolved. Nevertheless, the convergence of biosensors, molecular biology, AI-assisted modelling, semiconductor engineering and computational neuroscience may eventually allow entirely new approaches to dynamic biological monitoring.

Marta Amorim & Miguel Estrada