Virus research launchpad · Solana mainnet
Launch a coin on Meteora with your wallet, pick its species, and get a genomic research card from real NCBI sequences. 2.5% flat fee, migration to DAMM v2 at a $30k market cap. Live market and trades for every virus.
Every coin launched on Virus Research is bound to a virus species. The research worker fetches a real window of that species' reference genome from NCBI and computes a reproducible set of sequence statistics: the research card. No model is required for the baseline; the Viro1 genomic language models from Living Models can be plugged in for learned signal (research licence only).
species attribute, a scientific name or a common name). 27 catalogued virus species, every taxid verified against NCBI Taxonomy.kmer backend computes composition, CpG, entropy, GC profile, ORFs and low-complexity windows in a fraction of a second. The optional viro1 backend adds embeddings and masked pseudo-likelihood.Share of guanine and cytosine among the bases. Virus genomes range from ~33% (influenza) to ~56% (herpesviruses); gene-rich regions tend to be GC-richer than repeats.
How often the dinucleotide CG appears versus what the C and G frequencies predict. Values well below 1 indicate CpG depletion by methylation; near 1 means no depletion.
Shannon entropy in bits of the distribution of words of length k (1, 2, 3 and 6). Higher is more diverse sequence; low values point to repeats and low complexity.
GC content in 32 consecutive windows along the sequence, shown as a line and as the 3D relief. Peaks and dips hint at isochores, genes and repeats.
Stretches from an ATG start codon to a stop codon on either strand, 90 nt or longer. The longest ones are candidates for protein-coding segments.
64-base windows whose single-base entropy falls under 1.2 bits: poly-A tracts, microsatellites and similar repeats.
In comparisons, exact 12-mers present in both sequences. A high Jaccard points to homology or shared repeats; cosine similarity at k≤4 only reflects composition.
With model access: mean embedding, its norm and the masked pseudo-negative-log-likelihood of equally spaced positions. Lower means the model finds the sequence more expected.
esearch, esummary, efetch); every card records the accession, coordinates and the SHA-256 of the sequence.| Token | Species | Source | Status | Sequence | Registered |
|---|
No viruses match.
Loading…
No species match.