An Optimization Scheme Based on the Simulated Annealing Algorithm for In situ DNA Microarray Synthesis

DNA microarray synthesis enables the large-scale and precise generation of DNA sequences for genomic research, data storage, and synthetic biology. However, the order of nucleotide addition significantly affects synthesis efficiency and accuracy. This study aims to model DNA microarray in situ synthesis as a traveling salesman problem (TSP) and to develop an optimized synthesis strategy.

A mathematical model for in situ microarray synthesis was established, and both greedy algorithms and a simulated annealing algorithm were applied to optimize the nucleotide addition order. The performance of these approaches was evaluated by comparing the number of synthesis cycles required at different sequence scales, ranging from 10 × 10 nt to 10000 × 120 nt arrays.

The optimized synthesis schemes effectively reduced the total number of synthesis cycles. At the 10 × 10 nt scale, simulated annealing reduced cycles by 40.65% compared to the traditional scheme and by 8.52% compared to the greedy algorithm. At larger scales (100 × 100 nt to 10000 × 120 nt), cycle reductions ranged from 33.80% to 37.26%, with simulated annealing outperforming the greedy algorithm by 2.68% to 3.42%. These reductions translated into significant savings in synthesis time, reagent consumption, and overall cost.

The simulated annealing-based optimization strategy demonstrates clear advantages in improving DNA microarray synthesis efficiency while reducing material usage and waste, thereby enhancing cost-effectiveness. Such improvements offer practical benefits for applications, including gene editing, drug development, and DNA data storage.