Uploaded August 2026 | Updated September 2026, 2 weeks ago
The DNAMIC project, part of the European Pathfinder program for DNA-based digital data storage, aims to develop an autonomous, end-to-end DNA data storage solution based on a microfactory. Long-term data archiving is our primary use case, and we use the OLOS/DLCM system (olos.swiss), which complies with the Open Archival Information System (OAIS - ISO 14721) reference model, to manage data ingestion, preservation, and access. We are currently developing a DNA connector, integrated with OLOS/DLCM, that serves as an interface with the microfactory. This connector comprises (1) the codec, which converts the binary representation of the Archival Information Package (AIP) into DNA and vice versa, and (2) the state machine that controls the microfactory modules through API calls to trsynthesis, PCR, and sequencing.
Since the microfactory is designed as a modular architecture, the OLOS/DLCM DNA connector must be technology-independent for synthesis and sequencing. This means that certain quality controls must be available and provided to the codec, which must be able to adapt to different error rates. Our approach is to structure oligonucleotides with a hierarchy of parts, which enables a multi-step correction system. For instance, the coding scheme divides the files into distinct logical blocks, allowing us to decode them independently with the level of precision appropriate to the error rate. Synchronization tags are also available if needed, and two types of error correction codes are included. For the clustering step, which is critical to decoding, we developed a custom clustering algorithm based on k-mers embedding and a state-of-the-art approximate nearest-neighbor method, which we implemented on GPU for greater efficiency. If the error rate is low, a minimal amount of information is used for decoding. For silicon-based synthesis, and state-of-the-art nanopore base callers, data is decoded efficiently. For alternative synthesis solutions with higher error rates, such as photolithographic synthesis, for example, we exploit the many parts we have introduced into the oligonucleotides. The codec supports nucleotide resynchronization when the deletion (or insertion) rate is significantly high. When this is not sufficient to recover the data, error correction codes combined with error detection codes are used to correct the oligonucleotides through combinatorial trials. This method is obviously less effective, but it has the advantage of being able to recover information even when the error rate is high. This approach has been successfully tested on data with different error rates. The next steps will be to further optimize processing on dedicated infrastructure.
This is a presentation from the 2026 Storage and Computing with DNA Conference.
· Learn More about the SNIA DNA Data Storage Alliance: snia.org/groups/snia-dna-technology-affiliate
· SNIA Educational Library: snia.org/library
· X: twitter.com/SNIA
· LinkedIn: linkedin.com/company/snia
Presented by
Pierre-Yves Burgi & Michaël El Kharoubi
The DNAMIC project, part of the European Pathfinder program for DNA-based digital data storage, aims to develop an autonomous, end-to-end DNA data storage solution based on a microfactory. Long-term data archiving is our primary use case, and we use the OLOS/DLCM system (olos.swiss), which complies with the Open Archival Information System (OAIS - ISO 14721) reference model, to manage data ingestion, preservation, and access. We are currently developing a DNA connector, integrated with OLOS/DLCM, that serves as an interface with the microfactory. This connector comprises (1) the codec, which converts the binary representation of the Archival Information Package (AIP) into DNA and vice versa, and (2) the state machine that controls the microfactory modules through API calls to trsynthesis, PCR, and sequencing.
Since the microfactory is designed as a modular architecture, the OLOS/DLCM DNA connector must be technology-independent for synthesis and sequencing. This means that certain quality controls must be available and provided to the codec, which must be able to adapt to different error rates. Our approach is to structure oligonucleotides with a hierarchy of parts, which enables a multi-step correction system. For instance, the coding scheme divides the files into distinct logical blocks, allowing us to decode them independently with the level of precision appropriate to the error rate. Synchronization tags are also available if needed, and two types of error correction codes are included. For the clustering step, which is critical to decoding, we developed a custom clustering algorithm based on k-mers embedding and a state-of-the-art approximate nearest-neighbor method, which we implemented on GPU for greater efficiency. If the error rate is low, a minimal amount of information is used for decoding. For silicon-based synthesis, and state-of-the-art nanopore base callers, data is decoded efficiently. For alternative synthesis solutions with higher error rates, such as photolithographic synthesis, for example, we exploit the many parts we have introduced into the oligonucleotides. The codec supports nucleotide resynchronization when the deletion (or insertion) rate is significantly high. When this is not sufficient to recover the data, error correction codes combined with error detection codes are used to correct the oligonucleotides through combinatorial trials. This method is obviously less effective, but it has the advantage of being able to recover information even when the error rate is high. This approach has been successfully tested on data with different error rates. The next steps will be to further optimize processing on dedicated infrastructure.
This is a presentation from the 2026 Storage and Computing with DNA Conference.
· Learn More about the SNIA DNA Data Storage Alliance: snia.org/groups/snia-dna-technology-affiliate
· SNIA Educational Library: snia.org/library
· X: twitter.com/SNIA
· LinkedIn: linkedin.com/company/snia
Presented by
Pierre-Yves Burgi & Michaël El Kharoubi
