Approaches to data analysis Sample Clauses

Approaches to data analysis. The first step in next-generation sequencing analysis is the successful alignment of short reads to a reference genome. This is particularly challenging in bisulfite- sequencing (BS-seq) mapping, as the methylation status is inferred by comparing the reads to an unmodified reference sequence and the complexity of the libraries is reduced, due to the occurrence of cytosine to thymine conversions. Because of the complexity of these issues, a number of ad hoc alignment tools have been developed. Unbiased alignment tools, such as Bismark and BRAT convert to thymine any residual cytosine in the BS-seq reads and all the cytosines in the reference sequence (Xxxxxx, Xxxxx et al. 2010; Xxxxxxx and Xxxxxxx 2011). Thus, read mapping is not influenced by methylation status and converted reads show an exact match with the converted reference. Methylation-‘aware’ alignment tools consider both cytosines and thymines as potential matches to a genomic cytosine. Although these softwares occasionally favour the alignment of methylated reads, their mapping efficiency is higher than that of unbiased tools (Xxxxxxx, Xxxxx et al. 2012). A popular methylation-aware program is Novomethyl (Novocraft Technologies Sdn BhD), which aligns bisulfite reads to a reference genome through an iterative process, whereby an exact match is initially sought and the mismatch threshold is gradually increased in the subsequent rounds of analysis. Once an alignment has been generated a number of checks need to be implemented before the analysis can be completed. These include defining minimum read depths, estimating false discovery rate or cytosine conversion errors, integrating genomic variation data to efficiently detect C/T polymorphisms. Upon completion of these steps, the methylation status of cytosines can be inferred, and the ratio of unmethylated versus methylated CpG sites is determined.