In order to develop powerful computational predictors for identifying the biological features or attributes of DNAs, one of the most challenging problems is to find a suitable approach to effectively represent the DNA sequences. To facilitate the studies of DNAs and nucleotides, we developed a Python package called representations of DNAs (repDNA) for generating the widely used features reflecting the physicochemical properties and sequence-order effects of DNAs and nucleotides. There are 3 feature groups composed of 15 features. The first group calculates 3 nucleic acid composition features describing the local sequence information by means of kmers; the second group calculates 6 autocorrelation features describing the level of correlation between two oligonucleotides along a DNA sequence in terms of their specific physicochemical properties; the third group calculates 6 pseudo nucleotide composition features, which can be used to represent a DNA sequence with a discrete model or vector yet still keep considerable sequence order information via the physicochemical properties of its constituent oligonucleotides. In addition, these features can be easily calculated based on both the built-in and user-defined properties via using repDNA.
Copyright@By Liu Lab, Beijing Institute of Technology.