Fig. 6: A computational model can help determine system parameters for DNA data storage.
a A synthesis pool was generated with Nseq = 10,000 total number of sequences, with normally distributed copy numbers with a mean of \(\bar n_{syn}\) = 108 and standard deviation σ = 3.2 × 107. The pool was simulated to store an average copy number \(\bar n_0\) = 100, followed by 20 cycles of PCR amplification with P = 0.95, and high-throughput sequencing with average sequencing coverage \(\bar n_r\) = 200. Sequence dropout (i.e., coverage of 0 for a given sequence) rates were quantified. b Sequence dropout percentage as a function of variable synthesis pool c.v. and variable mean sequencing coverage \(\bar n_r\). c Sequence dropout percentage as a function of variable mean copy number n0 and variable mean sequencing coverage \(\bar n_r\). d Percentage of recovered oligo and logical redundancy as a function of average oligo copy number. The mean sequencing coverage is 5. When a very low copy number of oligo is stored, more sequences drop out, and thus higher logical redundancy is needed. The reported dropout percentage was the average of 100 repeated simulations.
