Fig. 3

SSBlazer captures specific motif patterns of the SSB site context and determines the crucial role of CpG dinucleotide. a Integrated gradients analysis in S1 END-seq dataset reveals contribution scores of ground truth SSB site context. The individual CpG pattern is discontinuously distributed in the SSB site context, and the motifs of CCG and CGG are disfavored in SSB site context. The motif of ATCAAT also arises in the SSB site context in S1 END-seq dataset. b, d Correlation analysis demonstrated the association between CpG content and predicted SSB counts in 212 species using Pearson correlation test (S1 END-seq: r = 0.86, p = \(2.9 \times e^{-64}\), SSiNGLe-ILM: r = 0.80, p = \(2.5 \times e^{-49}\)). c Contribution scores in the SSiNGLe-ILM dataset discover an explicit motif of GGC in the center region of the SSB site sequence context