Cognimed is the research program of this site: computing the body from the genome. It treats development as a cognitive process — the genome read as a large model, the tissue as a neural cellular automaton, and the bioelectric field as the latent space in which anatomy is specified. The themes covered on these pages — the TERRA RNAs, the HOX clusters, the Polycomb complex, and the bioelectric layer — are its working parts.

The papers

Nine papers are published open-access on Zenodo (CC-BY 4.0). Author: Miles Bradford Jacobs, genetec.io.

  1. Cognitive Biology: From Systems to Somatic Networks, Tissue Cognition, Cognitive Morphogenesis, and Tissue Emotions — DOI 10.5281/zenodo.20722139
  2. Cognitive Biology: Perceptrons and Morphogen Primordia — DOI 10.5281/zenodo.21143761
  3. Cognitive Biology Across Phyla: One Bioelectric Operator and a Substrate Reader That Sets the Boundary — DOI 10.5281/zenodo.20746637
  4. Cognitive Biology and Organ Formation: Organs as Master-Transcription-Factor Heads Read From an Accessibility Code — DOI 10.5281/zenodo.20925727
  5. Cognitive Biology: The Inner Perceptron versus the Outer Perceptron — DOI 10.5281/zenodo.21143016
  6. Cognitive Biology: Computing the Embryo — DOI 10.5281/zenodo.21221930
  7. Cognitive Biology: The Human Model and the GWAS Scaffold — DOI 10.5281/zenodo.21796907
  8. Cognitive Biology: Differentiation Clocks, Organ Formation and the MLP — DOI 10.5281/zenodo.21322049
  9. Cognitive Biology: The Mouse Model and the Attention-Head Manifold — DOI 10.5281/zenodo.21791849
  10. Cognitive Biology: Assembling the Body by Relational Attractors — in preparation

Code and data

Every published result is reproducible from the open series repository: github.com/AlphaFanX/cognitive-biology (MIT for code, CC-BY for the papers). The model is trained on public spatial transcriptomic atlases (MOSTA, ZESTA, HESTA) and grounded in public regulatory databases (AlphaGenome, SEdb, ABC).

The model itself is published on Hugging Face: huggingface.co/jacobsme/cognimed-nca-lgm — the complete NCA+LGM parameter set: the W0 kernel modules, the genome kernel/adapter partition, the per-allele adapter tables, the head maps and conductances, every derived result table, and our dense 3D atlas reconstructions. The whole model is 37 MB — its parameters are read from the genome through public databases, not fitted at scale: a glass box, small enough to audit line by line.

Why it matters

A computational model of development faithful enough to be interrogated is the precondition for in-silico medicine: designing and validating interventions against the model before the wet lab. The long-range targets are partial epigenetic reprogramming and aging, regenerative medicine, and gene therapy on the now-open AAV platform.