Figure 3.

Error correction methods for the DNA database. (A) Reed-Solomon codes multiplied the data block with a pre-calculated encoding matrix, and the redundant information would be added at the end of the result (108). (B) Fountain codes separated data into many segments, selected them using a special distribution function and packaged them into many ‘droplets’. Unqualified sequences were excluded from the screening procedure. Droplets of good quality were used for oligonucleotide synthesis (16). (C) Overlapping was a simple but effective method to avoid errors. Repeated pieces of the sequence under a shifting pattern were included in different oligonucleotide strands to obtain available copies (For example, four-fold redundancy generated in the figure) (13). (D) Exclusive-or operations between any two information strands generated the third strand. Any two of them could restore the remaining strand (26).