Bottom line up front: Genomes can store information. Conserved noncoding elements can survive hundreds of millions of years. Under optimal conditions, a genetic information channel spanning a billion years could theoretically preserve roughly 2.5 megabytes of data. This channel deserves to be taken seriously.
A Neglected Channel
For decades, we’ve scanned the sky with radio telescopes, looking for narrowband signals. But is it possible that the message already arrived—not in electromagnetic waves, but in the genomes of living cells?
In 2007, Masaru Tomita’s group at Keio University did something considered “crazy” at the time: they encoded Einstein’s E=mc² and 1905 into the DNA of Bacillus subtilis. Tomita said something key: “Bacterial information could persist—potentially for an incredible million years.”
More importantly, the design: he used fourfold redundant storage. Even if DNA mutates, the data can still be recovered. Tomita also anticipated today’s question: “The question a lot of people ask is—who would read the bacteria?” His answer had a rare scientific romanticism: “If aliens discover it in the distant future, they would naturally have higher intelligence to quickly crack the code.”
In 2026, Frontiers in Astronomy and Space Sciences published a hypothesis paper titled “Directed Information Panspermia.” The author, Sh. M. Kocharyan, proposed a concrete framework: a civilization could introduce primitive life forms to a habitable planet while simultaneously encoding information into the genetic code of those life forms. The genetic code, beyond its biological function, could also serve a “hyperbiological” role—as an interstellar information medium. The receiver civilization—the descendants evolved from those life forms—would eventually evolve the ability to read it.
This is a theoretical hypothesis, not an observational discovery. But it provides an operational framework.
So How Much Could Theoretically Be Stored?
In our earlier conversation, we attempted an estimate of this channel’s information capacity. The setup:
- Carrier: A large, slow-reproducing organism (crocodile-like), with a per-generation mutation rate of about 10⁻⁹ and a generation time of about 20 years.
- Time: 10⁹ years, roughly 5 × 10⁷ generations.
- Redundancy: Assume a large number of independent lineages carrying the same information, allowing a decoder to correct individual lineage mutations through consensus.
- Selection pressure: The information is embedded in conserved noncoding regions, hitchhiking on the host’s essential regulatory sequences.
Under this setup, the per-site substitution probability is about 4.9%. With an optimal decoder capable of maximum-likelihood consensus across millions of independent copies, the consensus error rate drops to negligible levels. Effective encoding capacity approaches 2 bits per base (the full four-letter alphabet).
If 1% of the genome can be used to carry information (about 10⁷ bases, for a large vertebrate genome), the recoverable information is roughly:
10⁷ bases × 2 bits/base = 2 × 10⁷ bits ≈ 2.5 megabytes
This is where the 2.5 MB figure comes from. It requires a series of ideal conditions to hold simultaneously: independent redundant lineages, active maintenance mechanisms preventing information deletion, and a decoder that knows where to look. But it gives an order of magnitude. The channel is narrowband, but it isn’t zero.
Why This Channel Might Be Real
The Genetic Code Itself Is an Error-Correcting Code
The standard genetic code is not random. It was optimized by natural selection into an error-correcting code. The codon assignment pattern minimizes the impact of point mutations on protein chemical properties. Biochemically similar amino acids are assigned to codons sharing two nucleotides—even if a mutation occurs, the consequence is “mild.”
If the genetic code itself is an error-correcting code, then any “unnatural” sequence in a genome inherits this error-correction capability. This is the underlying architecture that lets information hitchhike.
Conserved Noncoding Elements Prove Deep-Time Possibility
Some conserved noncoding elements are completely identical across humans, rats, mice, chickens, and dogs—species that diverged 300 million years ago. This is empirical proof: if selection pressure maintains it, noncoding sequence can survive hundreds of millions of years.
A designed anchor would be embedded in such a region—hitchhiking on the host’s essential regulatory sequences. It would sit in a place that “should drift but somehow doesn’t.” Its conservation would exceed what natural selection can explain.
Interstellar Material Exchange Is Real
Even if the information exists, it needs to reach Earth. 3I/ATLAS provides observational motivation. JWST/MIRI detected methane directly on an interstellar object for the first time, finding an unexpectedly high CH₄-to-water ratio. SPHEREx detected broad organic C–H features.
These molecules themselves are not biosignatures. But 3I/ATLAS proved one thing: material from the galaxy’s ancient regions can travel across vast interstellar distances to reach our solar system. It provides a concrete diagnostic framework: timing of volatiles, isotope ratios, chirality, spatial distribution of organics.
Honest Caveats
This estimate depends on several heroic assumptions that need to be explicitly flagged:
Independence. You need a large number of reproductively isolated lineages, each carrying the information, for a billion years. This isn’t “a ton of biological material”—it’s an engineered and maintained ecosystem.
Active maintenance. If the information serves no function, it is evolutionarily invisible. Billion-year information persistence requires active maintenance mechanisms: a toxin-antitoxin system, a homing endonuclease, a CRISPR-style “immune” system that kills cells that lose the information.
The recognizability paradox. A “recognizable message” is, by definition, something selection would not preserve. If it looks like a message, selection eliminates it. If it survives, it looks like biology.
The physical bottleneck. Billion-year-old fossil DNA has long since degraded. The information must be continuously replicated—meaning living organisms, not fossils. Maintaining a specific message for a billion years requires continuous selection or maintenance.
Worth Starting to Look
If panspermia was directed, the most likely trace is not the signal itself, but a conserved element that shouldn’t exist—a noncoding sequence whose conservation has no evolutionary explanation.
This claim is testable and falsifiable. It doesn’t require anyone to announce aliens. It requires only that we get better at spotting statistical anomalies in genomes.
The first evidence of extraterrestrial intelligence may not be a radio signal—it may be a stretch of DNA that is too conserved to be natural. We now have the tools to start looking. Whether we find it in two years or two hundred is unknown. But the search is no longer science fiction.
Acknowledgment: The logical framework of this post emerged from a conversation about panspermia, genetic repair mechanisms, and information theory. Any errors in the derivation belong to the conversationalists, not to the cited research.