TRL 4 Demonstration of the Synthesis of 10,000+ Nucleotide Pair Constructs with Per-Base Error < 0.0001

This Phase I project will demonstrate the use of Firebird technology in &quot;dual use&quot; commercial processes to synthesize large DNA fragments for use in synthetic biology. The key Phase I metric is the self-assembly of ~500 DNA fragments, each ~ 50 nucleotides long, to give transformation-competent plasmids with ~10,000 base pairs. These constructs must, on average after Phase II is complete, have less than one error, and an overall yield greater that 5%. The state-of-the-art relies on rules taught in high school to describe the behavior of the DNA double helix (&quot;A pairs with T, G pairs with C&quot;). Unfortunately, these rules do not capture the diverse biophysics that real DNA actually displays. In fact, non-standard binding interactions, mismatches, hairpin formation, and a variety of other facts prevent autonomous self-assembly of more than a dozen or two single stranded DNA fragments to give duplexes having a size able to transform synthetic biology if done &quot;at cost&quot;. In informatics language, the information density of standard DNA is too low to encode, in the sequences themselves, the information needed for self-assembly in competition with undesired assemblies. Further, the rate of assembly slows with increasing numbers of fragments. For this reason, Merryman and Gibson placed limits of direct assembly at &quot;about a dozen&quot; fragments. Even this requires pre-annealing of duplexes and, as the multiplex grows, expensive purification of individual DNA fragments. Our approach uses 8 different nucleotides extracted from an Artificially Expanded Genetic Information System (AEGIS) to increase the information density of the fragments. This allows them to autonomously self-assemble to give ultra large constructs. We also exploit reagent innovations to eliminate truncated and mis-sequence impurities from those fragments, allowing them to be used without any purification steps. Finally, we use patented transliteration tools to convert 8-letter DNA constructs into standard 4-letter DNA. These tools will shrink the DNA read-write &quot;bandwidth gap&quot;, diminishing cost of large assemblies by &gt; two orders of magnitude. The Phase I tasks to meet this Phase I metric involve synthesis of DNA fragments from more stable components, testing of new capping strategies to eliminate truncated fragments by simple filtration, performing self-assembly with fragments synthesized by phosphoramidite chemistry with these innovation, transliterate the assemblies to give standard 4-letter DNA, benchmark the errors of those assemblies by deep sequencing, use patterns to assign errors to mistakes made in synthesis, transliteration, amplification, and sequencing. The results will feed back into our Oligarch design software to support improved design that gives fewer errors. As commercialization benefits, Firebird estimates cumulative sales revenues of $300 million or cumulative licensing revenues of $30 million during the first 10 years of commercialization. This estimate relies on &quot;classical&quot; market analyses, capitalization of biotech companies in this space that do not have the AEGIS technology, and the emergence over the past 18 months of new applications for large DNA constructs, including in vaccines. For example, DARPA has just awarded $90 million, to create a 6' x 6' x 6' box that, upon deployment, makes 1000 doses of &gt;1000 nucleotide constructs in three days.