An algorithm for identifying protein-coding sequences and pseudogenes to improve microbial genome annotation = 향상된 미생물 유전체 주석 처리를 위한 단백질 발현 유전자 및 위유전자 판별 알고리즘

Abstract

As more and more microbial genomes are being sequenced by next-generation sequencing technologies, automated genome annotation systems have become more important to process a vast amount of genome sequence information. The usage of multiple gene prediction programs warrants higher sensitivity, but this approach is ineffective in terms of specificity and accuracy because of increasing false positives. Furthermore, since many automated genome annotation systems do not distinguish pseudogenes from functional genes, manual curation is necessary to ensure high-quality annotation which is time-consuming and not always feasible. We developed GeneCuraid that aids the high confidence curation of protein-coding sequences and pseudogenes from genes predicted by automatic annotation tools. When the genome sequence of Escherichia coli K-12 MG1655 was used as a test data set, the algorithm improved specificity and accuracy of the annotation results were 24.33% and 91.90%, while maintaining sensitivity as high as 98.09%. Therefore, we expect that GeneCuraid algorithm would attain the high-quality genome annotation and help to reduce time and cost in manually determining correct protein-coding genes and pseudogenes.