lithoSpore — LTEE Reproduction and Portable Validation
7 LTEE modules reproducing Wiser, Barrick, Good, Blount, Burden, Tenaillon papers + Anderson disorder predictions. 75/75 checks, USB-deployable, cross-platform.
Domain
Long-Term Evolution Experiment (LTEE) reproduction — power-law fitness, mutation accumulation, allele trajectories, citrate innovation, BioBrick burden, 264-genome comparison, Anderson disorder predictions.
Repository: sporeGarden/lithoSpore
The Science Story
lithoSpore is the ecosystem’s first Targeted GuideStone — a self-contained artifact that reproduces key results from the LTEE, the longest-running evolutionary biology experiment in history. It targets the Barrick Lab at UT Austin, where the experiment continues past 75,000 generations with E. coli.
The artifact proves two things simultaneously: (1) the LTEE’s published results are computationally reproducible from public data, and (2) the ecoPrimals infrastructure — pure Rust, musl-static, zero dependencies — can carry scientific validation to any machine on the planet.
Headline Results
- 75/75 checks across 7 science modules — all PASS at Tier 2 (Rust)
- 6 published papers reproduced: Wiser 2013, Barrick 2009, Good 2017, Blount 2008/2012, Burden 2024, Tenaillon 2016
- Anderson disorder framework applied to LTEE fitness data — GOE/Poisson eigenvalue statistics validate the disorder analogy
- Single 5.1 MB binary (musl-static) replaces 7 module binaries + 11 shell scripts
- Cross-platform: Ubuntu, Alpine, Fedora, Debian, read-only FS, Windows (7.9 MB litho.exe)
- 108 unit tests + 16 integration tests + 15 chaos/fault-injection tests
Seven Science Modules
| # | Module | Paper | Method | Checks |
|---|---|---|---|---|
| 1 | ltee-fitness | Wiser 2013 (Science) | Nelder-Mead curve fitting, AIC/BIC model selection | 8/8 |
| 2 | ltee-mutations | Barrick 2009 (Nature) | Kimura fixation, Poisson neutral accumulation, Pearson molecular clock | 7/7 |
| 3 | ltee-alleles | Good 2017 (Nature) | Clonal interference dynamics, fixation probability, adaptation rate | 20/20 |
| 4 | ltee-citrate | Blount 2008/2012 (PNAS/Nature) | Citrate innovation cascade, potentiation, replay probabilities | 11/11 |
| 5 | ltee-biobricks | Burden 2024 (Nat Comms) | Metabolic burden, cost-benefit analysis | 6/6 |
| 6 | ltee-breseq | Tenaillon 2016 (Nature) | 264-genome comparison, parallel evolution significance | 16/16 |
| 7 | ltee-anderson | Anderson-QS framework | GOE/Poisson eigenvalue statistics, disorder mapping | 7/7 |
Researchers Reproduced
| Researcher | Department | Domain |
|---|---|---|
| Jeffrey Barrick | Molecular Biosciences, UT Austin | LTEE continuation, mutation dynamics |
| Richard Lenski | BEACON Center, MSU | LTEE founder, evolutionary biology |
| Michael Wiser | BEACON Center, MSU | Power-law fitness models |
| Benjamin Good | Physics, Stanford | Allele dynamics, clonal interference |
| Zachary Blount | BEACON Center, MSU | Citrate innovation |
| Olivier Tenaillon | INSERM, Paris | Genomic evolution |
What the Constraint Revealed
Building a single-binary artifact forced the consolidation of 7 module binaries into one litho CLI with argv[0] symlink detection. This eliminated subprocess spawning entirely — every module calls lib::run_validation() in-process. The same constraint demanded replacing all 11 shell scripts (fetch, assemble, chaos-test, deploy-test, validate) with pure Rust implementations using ureq, walkdir, blake3, and chrono.
External command calls (date, hostname, id -u) were replaced with filesystem reads and library calls, enabling genuine cross-platform support including Windows.
Architecture
litho-core ← shared library
├── validation/ tolerance framework, named tolerances
├── provenance/ liveSpore tracking, BLAKE3 anchoring
├── discovery/ env → UDS → TURN → standalone
├── stats/ shared statistics (pearson_r)
├── harness/ module skip/load/dispatch helpers
└── viz/ petalTongue DataBinding adapters
↑
├── ltee-fitness ← Module 1 (lib.rs + thin main.rs)
├── ltee-mutations ← Module 2
├── ltee-alleles ← Module 3
├── ltee-citrate ← Module 4
├── ltee-biobricks ← Module 5
├── ltee-breseq ← Module 6
├── ltee-anderson ← Module 7
└── ltee-cli ← 13 subcommands (validate, verify, fetch, assemble, ...)#![forbid(unsafe_code)] enforced workspace-wide across all 9 crates.
Cross-Spring Connections
- ← groundSpring: B1-B4 statistical methods — model fitting, fixation probability, AIC/BIC
- ← wetSpring: B7 breseq genome analysis — 264-genome parallel evolution
- ← hotSpring: Anderson localization spectral primitives — GOE/Poisson statistics
- ← neuralSpring: ML surrogate enrichment (additive, not blocking)
- → petalTongue: DataBinding adapters for all 7 modules, Interactive SceneGraph IPC
- → projectFOUNDATION: Threads 1, 2, 4, 7 — validated data sources and quantitative targets
Deployment Testing
| Environment | Method | Result |
|---|---|---|
| Ubuntu airgapped VM | libvirt + cloud-init | PASS — 75/75 checks |
| Ubuntu VPS VM | libvirt + cloud-init | PASS + liveSpore provenance |
| Alpine 3.20 chroot | chroot on host | PASS — musl libc portability |
| Read-only filesystem | remount ro | PASS — graceful degradation |
| Windows x86_64 | Wine 11 | PASS — self-test + validate + verify |
See lithoSpore guideStone Artifact for full artifact documentation.
baseCamp Papers
Paper 02 (Frozen Fossil Record, LTEE extension) — see baseCamp Science for full list.