Alignment Free Dissimilarities for Nucleosome Classification

TY - CHAP

T1 - Alignment Free Dissimilarities for Nucleosome Classification

AU - Lo Bosco, Giosue'

AU - Lo Bosco, Giosué

PY - 2016

Y1 - 2016

N2 - Epigenetic mechanisms such as nucleosome positioning, his- tone modications and DNA methylation play an important role in the regulation of cell type-specic gene activities, yet how epigenetic pat- terns are established and maintained remains poorly understood. Recent studies have shown a role of DNA sequences in recruitment of epige- netic regulators. For this reason, the use of more suitable similarities or dissimilarity between DNA sequences could help in the context of epi- genetic studies. In particular, alignment-free dissimilarities have already been successfully applied to identify distinct sequence features that are associated with epigenetic patterns and to predict epigenomic proles. In this work, we focalize the study on the problem of nucleosome classi- cation, providing a benchmark study of 6 alignment free dissimilarity measures between sequences, belonging to the categories of geometric- based, correlation-based, information-based and compression based. Their comparisons have been done versus an alignment based dissimilarity, by measuring the performance of several nearest neighbour classiers that incorporate each one the considered dissimilarities. Results computed on three dataset of nucleosome forming and inhibiting sequences, shows that among the alignment free dissimilarities, the geometric and correlation are the more suitable for the purpose of nucleosome classication, mak- ing them a more ecient alternative to the alignment-based similarity measures, which nevertheless are yet the preferred choice when dealing with sequence similarity measurements.

AB - Epigenetic mechanisms such as nucleosome positioning, his- tone modications and DNA methylation play an important role in the regulation of cell type-specic gene activities, yet how epigenetic pat- terns are established and maintained remains poorly understood. Recent studies have shown a role of DNA sequences in recruitment of epige- netic regulators. For this reason, the use of more suitable similarities or dissimilarity between DNA sequences could help in the context of epi- genetic studies. In particular, alignment-free dissimilarities have already been successfully applied to identify distinct sequence features that are associated with epigenetic patterns and to predict epigenomic proles. In this work, we focalize the study on the problem of nucleosome classi- cation, providing a benchmark study of 6 alignment free dissimilarity measures between sequences, belonging to the categories of geometric- based, correlation-based, information-based and compression based. Their comparisons have been done versus an alignment based dissimilarity, by measuring the performance of several nearest neighbour classiers that incorporate each one the considered dissimilarities. Results computed on three dataset of nucleosome forming and inhibiting sequences, shows that among the alignment free dissimilarities, the geometric and correlation are the more suitable for the purpose of nucleosome classication, mak- ing them a more ecient alternative to the alignment-based similarity measures, which nevertheless are yet the preferred choice when dealing with sequence similarity measurements.

KW - Alignment free DNA sequence dissimilarities

KW - Epigenetic

KW - Knn classifier

KW - L-tuples

KW - Nucleosome classification

KW - k-mers

UR - http://hdl.handle.net/10447/181678

UR - http://link.springer.com/chapter/10.1007/978-3-319-44332-4_9

M3 - Chapter

SN - 9783319443317

VL - 9874

T3 - LECTURE NOTES IN COMPUTER SCIENCE

SP - 114

EP - 128

BT - Computational Intelligence Methods for Bioinformatics and Biostatistics, 12th International Meeting, CIBB 2015, Naples, Italy, September 10-12, 2015, Revised Selected Papers

ER -