Project Evo-12K

The next twelve thousand years.

Artists and geneticists picturing the human body far past us. Each portrait is a hypothesis about the pressures that could shape it.

Trait atlas

One trait at a time.

Each portrait is a visual hypothesis. We take one feature at a time and separate it into the biology that would have to move together, the environment that could reward it, and the evidence we have now.

VIS-01 · sensory

Low-light vision

Useful night vision comes from a whole optical and neural system: aperture, lens transmission, rod density, retinal gain, and circadian timing. The cat-eye look is only its surface.

Speculative monochrome study of a human low-light eye and retinal architecture

Selection environment: Long twilight, subterranean habitats, low-power interiors, or worlds around dim M- and K-dwarf stars. The target is usable contrast at low photon counts.
Polygenic architecture: Pupil dynamics and orbital geometry must coordinate with rod-biased retina, rod-bipolar wiring, melanopsin signaling, and neural noise suppression. No single allele creates the system.
Comparative biology: Cats provide aperture and tapetum references; geckos show multifocal low-light optics; reindeer show seasonal changes in tapetal reflectance. Humans have no tapetum lucidum.
Biological cost: More photon capture can trade against daylight acuity, color resolution, glare recovery, and circadian stability. Bright environments would require behavioral or architectural shielding.
Candidate anchors: PAX6 · RHO · NRL · NR2E3 · PDE6B · GNAT1 · OPN4 · comparative feline tapetum

DERM-02 · barrier

Radiation and pigment

Skin color is only the visible output. A useful radiation phenotype couples melanin chemistry, melanosome distribution, DNA repair, barrier integrity, inflammation, and thermal balance.

Selection environment: High UV, thin atmospheres, F-star spectra, high altitude, or persistent glare from water, ice, and pale mineral surfaces.
Polygenic architecture: Eumelanin-to-pheomelanin balance, melanocyte signaling, melanosome maturation and transfer, keratinocyte response, and repair load must be modeled separately.
Comparative biology: Zebrafish slc24a5 established a conserved pigment mechanism; ASIP/MC1R switching patterns mammalian coats. Neither explains a blue or indigo human surface.
Biological cost: More absorption can raise heat load; less pigment raises UV injury risk. Vitamin-D ecology, clothing, microbiome, and habitat shielding remain part of the phenotype.
Candidate anchors: SLC24A5 rs1426654 · SLC45A2 rs16891982 · MC1R · ASIP · TYR · OCA2 · KITLG · FLG

PIL-03 · surface

Hair as a climate interface

Color, length, density, curl, shaft geometry, and growth cycle are distinct traits. Their value changes with scalp UV, humidity, airflow, cold, and maintenance cost.

Selection environment: Cold sealed habitats may reward insulation; high-radiance worlds may reward scalp shielding; humid heat changes the value of curl, density, and airflow.
Polygenic architecture: Follicle shape, keratin structure, anagen duration, shaft diameter, pigment, and ectodermal development form a multivariate surface rather than a single coat gene.
Comparative biology: FGF5 alters coat length across mammals; KRT71 affects curl and texture; hairless and seasonal-coat species expose growth-cycle tradeoffs. Human scalp hair remains developmentally distinct.
Biological cost: Insulation can impair heat rejection. Dense or long hair carries energy, hygiene, microbiome, and mechanical costs that the habitat must justify.
Candidate anchors: FGF5 · EDAR V370A · TCHH · KRT71 · HR · KITLG · MC1R

CRAN-04 · morphology

Facial architecture

A face is a coordinated 3D growth field. Orbit, brow, nose, cheek, jaw, and soft tissue cannot be assigned one gene each or inferred independently.

Selection environment: Orbital depth can alter glare exposure; nasal geometry interacts with airflow and humidity; jaw loading follows diet; sealed habitats change respiratory and social constraints.
Polygenic architecture: Multivariate models divide the face into correlated segments shaped by cranial neural crest regulation, development, allometry, sex, and age.
Comparative biology: Animal craniofacial diversity reveals developmental pathways, though cross-species looks do not transfer to humans. What carries over is the mechanism.
Biological cost: Changing one region can affect airway, bite, speech, vision, and development. Aesthetic endpoints are poor substitutes for functional phenotyping.
Candidate anchors: 203 lead GWAS signals · 63 facial segments · PAX1 · DCHS2 · RUNX2 · GLI3 · EDAR · TBX15/WARS2

SOMA-05 · body plan

Stature and proportion

Height looks simpler than it is. The target is a distribution of stature, limb proportion, bone geometry, and load tolerance.

Selection environment: Gravity, habitat clearance, locomotion distance, nutrition, thermal surface area, pregnancy, and injury risk determine whether a body plan is useful.
Polygenic architecture: A 5.4-million-person GWAS mapped 12,111 independent common associations. Prediction remains ancestry-dependent and must be coupled to development and environment.
Comparative biology: Long-bone growth pathways are conserved across vertebrates, while island and altitude ecologies show how body size changes under energy and terrain pressure.
Biological cost: Greater stature raises skeletal loading, circulation distance, calorie demand, and obstetric constraints. Smaller bodies alter reach, heat exchange, and power.
Candidate anchors: 12,111 associations · HMGA2 · GDF5 · NPR3 · ACAN · SHOX · growth plate pathways

NEUR-06 · performance

Cognitive phenotypes

Cognition has to be broken into measured domains. It cannot be read from a face or reduced to one score. The object of study is performance in a defined environment over time.

Selection environment: Long-duration missions, disrupted circadian cycles, multilingual groups, high automation, sensory scarcity, and chronic isolation may reward different cognitive profiles.
Polygenic architecture: Processing speed, working memory, attention control, sensory integration, learning, and sleep resilience have overlapping but non-identical polygenic and developmental bases.
Comparative biology: Animal models reveal conserved memory, sleep, and sensory pathways, but not human social or abstract cognition. That needs longitudinal phenotyping in people.
Biological cost: Current scores are population- and context-dependent. Pleiotropy, education, health, inequality, and ancestry portability make selection claims scientifically and ethically fragile.
Candidate anchors: 205 associated loci in a 269,867-person GWAS · distributed neurodevelopmental pathways · longitudinal tests · G×E

An open call to artists and geneticists.Request an invitation