The invention relates to the field of
computational biology and
bioinformatics, in particular to an
RNA (Ribonucleic Acid) sequence screening and
promoter function predicting and
screening method based on
machine learning, which comprises the following steps of: firstly, adopting a novel sequence
algorithm based on a sliding window and a position weight matrix, and combining technologies such as k-mer
frequency analysis and conservative motif recognition; candidate
RNA sequences are efficiently identified from genomic sequences. Then, an integrated prediction model fusing
deep learning and traditional
machine learning is established, and multi-dimensional information such as sequence features, structural features,
functional features and evolution features is extracted; and carrying out batch prediction on the candidate
RNA sequences, outputting a function intensity
score and providing credibility evaluation. And finally, screening an
RNA sequence with an optimal prediction result based on a multi-objective optimization strategy, and designing an experimental
verification scheme. According to the method, the prediction accuracy and the screening efficiency are improved, the experiment cost can be remarkably reduced, and an efficient
promoter screening tool is provided for the fields of
synthetic biology,
gene therapy, biological
pharmacy and the like.