DNA N6-methyladenine (6 mA) is one of the most vital epigenetic modifications and involved in controlling the various gene expression levels. With the avalanche of DNA sequences generated in numerous databases, the accurate identification of 6 mA plays an essential role for understanding molecular mechanisms. Because the experimental approaches are time-consuming and costly, it is desirable to develop a computation model for rapidly and accurately identifying 6 mA. To the best of our knowledge, we first proposed a computational model named i6mA-Fuse to predict 6 mA sites from the Rosaceae genomes, especially in Rosa chinensis and Fragaria vesca. We implemented the five encoding schemes, i.e., mononucleotide binary, dinucleotide binary, k-space spectral nucleotide, k-mer, and electron-ion interaction pseudo potential compositions, to build the five, single-encoding random forest (RF) models. The i6mA-Fuse uses a linear regression model to combine the predicted probability scores of the five, single encoding-based RF models. The resultant species-specific i6mA-Fuse achieved remarkably high performances with AUCs of 0.982 and 0.978 and with MCCs of 0.869 and 0.858 on the independent datasets of Rosa chinensis and Fragaria vesca, respectively. In the F. vesca-specific i6mA-Fuse, the MBE and EIIP contributed to 75% and 25% of the total prediction; in the R. chinensis-specific i6mA-Fuse, Kmer, MBE, and EIIP contribute to 15%, 65%, and 20% of the total prediction. To assist high-throughput prediction for DNA 6 mA identification, the i6mA-Fuse is publicly accessible at https://kurata14.bio.kyutech.ac.jp/i6mA-Fuse/.

中文翻译：

---

i6mA-Fuse：通过融合多个特征表示，对蔷薇科基因组中的DNA 6 mA位点进行改进而可靠的预测。

DNA N6-甲基腺嘌呤（6 mA）是最重要的表观遗传修饰之一，涉及控制各种基因表达水平。随着大量数据库中产生的DNA序列雪崩，对6 mA的准确识别对于理解分子机制起着至关重要的作用。由于实验方法既耗时又昂贵，因此需要开发一种计算模型来快速准确地识别6 mA。据我们所知，我们首先提出了一个名为i6mA-Fuse的计算模型，可以从蔷薇科的基因组中预测6 mA的位点，特别是在蔷薇和草莓中。我们实施了五种编码方案，即单核苷酸二元，二核苷酸二元，k-空间光谱核苷酸，k-mer和电子-离子相互作用假电位组成，以构建这五种，单编码随机森林（RF）模型。i6mA-Fuse使用线性回归模型来组合五个基于单个编码的RF模型的预测概率分数。在Rosa chinensis和Fragaria vesca的独立数据集上，所得的特定于物种的i6mA-Fuse分别以0.982和0.978的AUC和0.869和0.858的MCC取得了卓越的性能。在vesca特有的i6mA保险丝中，MBE和EIIP分别占总预测的75％和25％。在中华民国特定的i6mA-Fuse中，Kmer，MBE和EIIP占总预测的15％，65％和20％。为了帮助进行DNA 6 mA鉴定的高通量预测，可在https://kurata14.bio.kyutech.ac.jp/i6mA-Fuse/上公开访问i6mA-Fuse。i6mA-Fuse使用线性回归模型来组合五个基于单个编码的RF模型的预测概率分数。在Rosa chinensis和Fragaria vesca的独立数据集上，所得的特定于物种的i6mA-Fuse分别以0.982和0.978的AUC和0.869和0.858的MCC取得了卓越的性能。在vesca特有的i6mA保险丝中，MBE和EIIP分别占总预测的75％和25％。在中华民国特定的i6mA-Fuse中，Kmer，MBE和EIIP占总预测的15％，65％和20％。为了帮助进行DNA 6 mA鉴定的高通量预测，可在https://kurata14.bio.kyutech.ac.jp/i6mA-Fuse/上公开访问i6mA-Fuse。i6mA-Fuse使用线性回归模型来组合五个基于单个编码的RF模型的预测概率分数。在Rosa chinensis和Fragaria vesca的独立数据集上，所得的特定于物种的i6mA-Fuse分别以0.982和0.978的AUC和0.869和0.858的MCC取得了卓越的性能。在vesca特有的i6mA保险丝中，MBE和EIIP分别占总预测的75％和25％。在中华民国特定的i6mA-Fuse中，Kmer，MBE和EIIP占总预测的15％，65％和20％。为了帮助进行DNA 6 mA鉴定的高通量预测，可在https://kurata14.bio.kyutech.ac.jp/i6mA-Fuse/上公开访问i6mA-Fuse。在Rosa chinensis和Fragaria vesca的独立数据集上，所得的特定于物种的i6mA-Fuse分别以0.982和0.978的AUC和0.869和0.858的MCC取得了卓越的性能。在vesca特有的i6mA保险丝中，MBE和EIIP分别占总预测的75％和25％。在中华民国特定的i6mA-Fuse中，Kmer，MBE和EIIP占总预测的15％，65％和20％。为了帮助进行DNA 6 mA鉴定的高通量预测，可在https://kurata14.bio.kyutech.ac.jp/i6mA-Fuse/上公开访问i6mA-Fuse。在Rosa chinensis和Fragaria vesca的独立数据集上，所得的特定于物种的i6mA-Fuse分别以0.982和0.978的AUC和0.869和0.858的MCC取得了卓越的性能。在vesca特有的i6mA保险丝中，MBE和EIIP分别占总预测的75％和25％。在中华民国特定的i6mA-Fuse中，Kmer，MBE和EIIP占总预测的15％，65％和20％。为了帮助进行DNA 6 mA鉴定的高通量预测，可在https://kurata14.bio.kyutech.ac.jp/i6mA-Fuse/上公开访问i6mA-Fuse。MBE和EIIP分别占总预测的75％和25％；在中华民国特定的i6mA-Fuse中，Kmer，MBE和EIIP占总预测的15％，65％和20％。为了帮助进行DNA 6 mA鉴定的高通量预测，可在https://kurata14.bio.kyutech.ac.jp/i6mA-Fuse/上公开访问i6mA-Fuse。MBE和EIIP分别占总预测的75％和25％；在中华民国特定的i6mA-Fuse中，Kmer，MBE和EIIP占总预测的15％，65％和20％。为了帮助进行DNA 6 mA鉴定的高通量预测，可在https://kurata14.bio.kyutech.ac.jp/i6mA-Fuse/上公开访问i6mA-Fuse。