DNA primase is an essential enzyme that synthesizes short RNA primers on specific DNA sequences. These RNA primers are elongated by DNA polymerase to form Okazaki fragments on the lagging DNA strand. It is therefore reasonable to as-sume that the binding of DNA primase on a genome marks the start sites of the Okazaki fragments. It has long been known that the frequency of the occurrence of primase trinucleotide recognition on a genome sequence has no influence on the size of the Okazaki fragments. The unresolved enigma that we address in this study is therefore why some, but not all, primase-DNA recognition sequences (PDRSs) become Okazaki fragment start sites. To this end, we applied machine-learning algo-rithms to analyze a massive amount of data obtained from protein-DNA binding microarrays (PBM) with the aim of identi-fying the essential elements on DNA that are needed for the binding of bacteriophage T7 primase. A PBM data learning al-gorithm enabled the prediction of binding values of T7 primase for any given DNA sequence with unprecedented accuracy and flexibility. On the basis of the principles learned about DNA-primase binding, we generated novel DNA sequences with improved binding of T7 primase and improved RNA primer synthesis, as validated experimentally.

中文翻译：

---

通过使用数据驱动的方法来重新配置primase DNA识别序列

DNA引发酶是在特定DNA序列上合成短RNA引物的基本酶。这些RNA引物被DNA聚合酶拉长，在落后的DNA链上形成Okazaki片段。因此，可以合理地假设DNA primase在基因组上的结合标志着冈崎片段的起始位点。早就知道，在基因组序列上发生三联酶三核苷酸识别的频率对Okazaki片段的大小没有影响。因此，我们在这项研究中要解决的未解决的谜团就是为什么某些（但不是全部）引发酶DNA识别序列（PDRS）成为冈崎片段起始位点的原因。为此，我们应用了机器学习算法来分析从蛋白质-DNA结合微阵列（PBM）获得的大量数据，目的是识别结合噬菌体T7 primase所需的DNA基本元素。PBM数据学习算法能够以前所未有的准确性和灵活性预测任何给定DNA序列的T7 primase结合值。根据有关DNA引发酶结合的原理，我们通过实验验证了产生具有改善的T7引发酶结合能力和RNA引物合成能力的新型DNA序列。