Hybrid Error-Resilient DNA Data Storage: Bias-Aware Coding, Enzymatic Repair Simulation, and Microcapsule-Based Random Access
Md, . Mehedi Hasan
Abstract
DNA has emerged as a promising medium for long-term archival data storage due to its exceptional information density, durability, and passive energy requirements. However, practical deployment remains constrained by three fundamental challenges: stochastic synthesis bias, chemical degradation over time, and limited scalability of random access. This thesis proposes a Hybrid Error-Resilient DNA Storage Framework that integrates computational, biochemical, and architectural solutions to address these limitations holistically. First, a bias-aware adaptive coding scheme is introduced, which models synthesis and PCR dropout as probabilistic processes and dynamically assigns logical redundancy based on predicted sequence fragility. Unlike conventional uniform redundancy approaches, the proposed method reduces sequencing coverage requirements while preserving decoding reliability. Second, a Markov-chain-based enzymatic repair model is developed to simulate long-term molecular decay and restoration using a multi-enzyme repair cocktail consisting of APE1, Bst polymerase, and Taq ligase. The model demonstrates a significant extension of data recoverability horizons under accelerated aging conditions. Third, the thesis evaluates thermoresponsive microcapsule based random access, enabling repeated, low-bias retrieval through thermoconfined PCR while preventing destructive consumption of the archival pool. Comprehensive simulations indicate a reduction of sequencing coverage by over 70% and an effective extension of archive longevity by nearly threefold compared to conventional DNA storage pipelines. Additionally, a sustainability analysis highlights substantial reductions in energy consumption, carbon emissions, and electronic waste relative to magnetic tape and hard disk-based archival systems. The results establish DNA storage as a viable candidate for future ultra-long term, sustainable digital preservation.
Permanent link — send this address to others:https://lib.su.edu.bd/thesis/123456789/2610