A Fast Statistical Biological Sequence Compressor for Pattern Discovery

This paper introduces a novel algorithm for biological sequence compression that makes use of both statistical properties and repetition within sequences by maintaining a panel of experts to predict the probability distribution of the next symbol in the sequence to be encoded. Expert probabilities are combined to obtain the final distribution. Experiments show that our algorithm outperforms existing compressors on typical DNA and protein sequence datasets while maintaining a practical running time. The resulting information sequence provides insight for further study of the biological sequence. We demonstrate a number pattern discovery tasks using our model.