Skip to content

000specific/GIGANTIC

GIGANTIC

GIGANTIC is a modular framework for AI-assisted comparative genomics and phylogenomics. It is designed to be cloned from GitHub, copied into a renamed project directory, and then operated through AI assistant sessions rooted at that renamed project directory.

The mantra: you (the user) guide an AI; the AI does the work. GIGANTIC's structure, conventions, and documentation are built around that model.

GIGANTIC use is built around a given project implemented through human-AI chat. It provides a framework of subprojects containing automated bioinformatic and phylogenetic workflows under the hood. It includes extensive documentation at every turn - READMEs for humans and AI_GUIDEs for AIs. For the non-expert user, it is intended as a wandering yet guided - often surprising & hopefully more fun than frustrating - research conversation to explore species, genomes, ideas, and data with an AI in chat. One where the human is guided and guides - and the AI acts. GIGANTIC floats you along in conversation - but with an exosuit grounded in sequence exploration - built to provide rigorous tools and methods in phylogenomics - and less excitingly but perhaps most critically - on-demand and behind the scenes documentation of chat as a record of research development and implementation, providing provenance in support of attribution, transparency, reproducibility, and publication. In this sense, a GIGANTIC project serves as a digital research notebook for comparative genomic and phylogenetic work - intended in the end to be archived and maintained no differently than a lab notebook at the bench.

Species for the demo are below - species set is species42. It is built for discovery of gene and genome evolution within a comparative framework of deuterostomes. It leverages deep functional knowledge in human, mouse and to a lesser extent zebrafish - and in the worm and fly models as outgroups. Hippo & splattering of cetaceans - plus octopus - fill things out. Entry requirements into species42 included a public high-quality annotated genome.

Demo species set (species42, 42 species, deuterostome-focused):

Phylum Species Common name Genome
Chordata Homo sapiens human GCF_000001405.40
Chordata Mus musculus house mouse GCF_000001635.27
Chordata Danio rerio zebrafish GCF_049306965.1
Chordata Balaenoptera musculus blue whale GCF_009873245.2
Chordata Megaptera novaeangliae humpback whale GCA_041834305.1
Chordata Balaenoptera borealis sei whale GCF_965194805.1
Chordata Orcinus orca killer whale GCF_937001465.1
Chordata Tursiops truncatus common bottlenose dolphin GCF_011762595.2
Chordata Phocoena phocoena harbor porpoise GCF_963924675.1
Chordata Hippopotamus amphibius kiboko hippo GCF_030028045.1
Chordata (Cephalochordata) Branchiostoma floridae Floridian lancelet GCF_000003815.2
Chordata (Cephalochordata) Branchiostoma lanceolatum Mediterranean amphioxus GCF_035083965.1
Chordata (Urochordata) Ascidia mentula sea squirt GCF_947561715.1
Chordata (Urochordata) Botryllus schlosseri star tunicate GCF_051294905.1
Chordata (Urochordata) Corella eumyota orange-tipped sea squirt GCF_963082875.1
Chordata (Urochordata) Phallusia mammillata white sea squirt GCF_965637545.1
Chordata (Urochordata) Styela clava rough sea squirt GCF_964204865.1
Chordata (Urochordata) Aplidium turbinatum colonial sea squirt GCF_918807975.1
Chordata (Urochordata) Ciona intestinalis vase tunicate GCF_018327825.1
Chordata (Urochordata) Clavelina lepadiformis light-bulb sea squirt GCF_947623445.1
Chordata (Urochordata) Oikopleura dioica larvacean GCF_907165135.1
Chordata (Urochordata) Thalia democratica salp GCF_965202585.1
Echinodermata Antedon bifida rosy feather star GCF_963402885.1
Echinodermata Asterias rubens European sea star GCF_902459465.1
Echinodermata Lytechinus pictus Cyclins Nobel Prize painted sea urchin GCF_037042905.1
Echinodermata Amphiura filiformis brittlestar GCF_039555335.1
Echinodermata Antedon mediterranea Mediterranean feather star GCF_964355755.1
Echinodermata Apostichopus japonicus Japanese sea cucumber GCF_037975245.1
Echinodermata Asterias amurensis Northern Pacific sea star GCF_032118995.1
Echinodermata Diadema setosum long-spined sea urchin GCF_964275005.2
Echinodermata Eupentacta fraudatrix sea cucumber GCA_056552365.1
Echinodermata Holothuria leucospilota black sea cucumber GCA_029531755.1
Echinodermata Lytechinus variegatus green sea urchin GCF_018143015.1
Echinodermata Mesocentrotus franciscanus red sea urchin GCA_037976565.1
Echinodermata Patiria miniata bat sea star GCF_015706575.1
Echinodermata Strongylocentrotus purpuratus purple sea urchin GCF_000002235.5
Hemichordata Glandiceps talaboti enteropneust acorn worm GCF_964340395.1
Hemichordata Ptychodera flava enteropneust acorn worm GCF_041260155.1
Hemichordata Schizocardium californicum enteropneust acorn worm figshare:25816303
Nematoda Caenorhabditis elegans roundworm GCF_000002985.6
Arthropoda Drosophila melanogaster fruit fly GCF_000001215.4
Mollusca Octopus bimaculoides California two-spot octopus GCF_001194135.2

Getting started — three steps

1. Clone the repository

git clone https://github.com/000generic/GIGANTIC.git

2. Copy and rename the project template

gigantic_project-COPYME/ (inside the cloned GIGANTIC/) is the project template. Copy it to a working location outside the GIGANTIC/ clone, and rename it to something specific to your research project:

cp -r GIGANTIC/gigantic_project-COPYME /path/to/work/gigantic_project-cephalopod_evolution

Project naming convention: gigantic_project-<your_project_name>. Examples:

  • gigantic_project-cephalopod_evolution
  • gigantic_project-early_animal_phylogenomics
  • gigantic_project-mollusc_neural_genes

The renamed directory is now your project. It is self-contained: every input, workflow, output, and documentation file lives inside it. You can archive the whole directory as the complete record of your project.

3. Start a fresh, naive AI assistant session rooted at the renamed project

Open a new AI assistant session (Claude Code, Cursor, ChatGPT, Gemini, or any other codebase-aware AI assistant) and root it at the renamed project directory, not at the cloned GIGANTIC/ directory.

For example, if you renamed your project to gigantic_project-cephalopod_evolution, your AI session should treat that directory as its working root.

All subsequent project work happens through AI sessions rooted at your renamed project directory. The GIGANTIC/ clone is the framework; your renamed project directory is where you actually do science.


Why this structure

  • GIGANTIC/ is the framework that gets versioned on GitHub. It ships; it doesn't run.
  • Your renamed project is what runs. It is the canonical record of your research project — fully reproducible, fully archivable, fully separable from any other project that uses GIGANTIC.
  • AI sessions rooted at the renamed project stay focused on your project scope and don't have to mentally subtract framework-development context from every response.

Where to find more

  • Project template setup, conventions, and AI guidance: inside gigantic_project-COPYME/ — specifically README.md, AI_GUIDE.md, and CLAUDE.md in that directory. Their content travels with your renamed copy.
  • AI-assistant-facing guidance for the framework itself: AI_GUIDE.md in this directory.
  • Behavior and posture for research-grade work (full documentation, transparency, archival, replication): see CLAUDE.md and AI_BEHAVIOR.md inside gigantic_project-COPYME/ (these travel with each renamed project copy — research-grade posture belongs at the project level, not at the framework level).
  • Standard repository files: LICENSE, CITATION.cff, CONTRIBUTING.md, CODE_OF_CONDUCT.md.

Detailed documentation of subprojects, workflows, conventions, and AI-assistant behavior lives inside gigantic_project-COPYME/ and travels with each renamed project copy. This top-level README is intentionally a minimal landing page.

About

GIGANTIC: Genome Integration and Gene Analysis across Numerous Topology-Interrogated Clades

Resources

License

Code of conduct

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors