From Data to Codewords
Information is divided into symbols. Each symbol is mapped to a fixed-length codeword a bit string designed with uniform properties such as a constant number of ones and zeros. This ensures:- Simplified error handling
- High signal-to-noise ratios during retrieval
- Consistent density across the archive
Combinatorial Identifiers
Rather than creating each codeword base-by-base, xDNA constructs molecules called identifiers. An identifier is built by assembling prefabricated DNA components layer by layer.Layer-by-layer assembly
Each layer contains multiple possible DNA components. One component is
chosen from each layer and ligated together the result is a unique sequence
that encodes a specific part of the dataset.
Exponential address space
With only a few hundred components divided into layers, it is possible to
assemble trillions of distinct identifiers exponential expansion with
simple, cost-effective chemistry.
Trie-like Structure
The identifiers naturally form a trie a tree-like data structure where each path from root to leaf corresponds to a sequence of components that encodes a codeword. This organization makes it possible to later traverse and query the archive using biochemical operations rather than digital scanning.The trie structure is what makes in-DNA compute possible. It turns the archive
into an associative memory that can be searched and filtered chemically,
without reading the full dataset.
Writing Process
1
Preprocess & encode
Digital data is preprocessed and mapped into codewords.
2
Assign identifiers
Each bit in a codeword is assigned to a unique identifier within the
combinatorial address space.
3
Assemble molecules
Identifiers are assembled from DNA components using chemical ligation in
parallel reactions.
4
Archive the library
The resulting DNA library contains millions of identifiers, each
representing a portion of the dataset stable, inert, and ready for storage
or computation.