SMAP: a streamlined methylation analysis pipeline for bisulfite sequencing

Abstract

Background: DNA methylation has important roles in the regulation of gene expression and cellular specification. Reduced representation bisulfite sequencing (RRBS) has prevailed in methylation studies due to its cost-effectiveness and single-base resolution. The rapid accumulation of RRBS data demands well designed analytical tools.

Findings: To streamline the data processing of DNA methylation from multiple RRBS samples, we present a flexible pipeline named SMAP, whose features include: (i) handling of single-and/or paired-end diverse bisulfite sequencing data with reduced false-positive rates in differentially methylated regions; (ii) detection of allele-specific methylation events with improved algorithms; (iii) a built-in pipeline for detection of novel single nucleotide polymorphisms (SNPs); (iv) support of multiple user-defined restriction enzymes; (v) conduction of all methylation analyses in a single-step operation when well configured.

Conclusions: Simulation and experimental data validated the high accuracy of SMAP for SNP detection and methylation identification. Most analyses required in methylation studies (such as estimation of methylation levels, differentially methylated cytosine groups, and allele-specific methylation regions) can be executed readily with SMAP. All raw data from diverse samples could be processed in parallel and 'packetized' streams. A simple user guide to the methylation applications is also provided.

Keywords: Allele-specific DNA methylation (ASM); Differentially methylated region (DMR); Reduced representation bisulfite sequencing (RRBS).

Fig. 1

Brief workflow of SMAP. SMAP filters sequenced raw data and produces ‘clean data’. Clean data is then mapped onto the pretreated reference genome. Finally, SNPs, ASM, DMRs and DMCs are detected and reports created. Existing tools are in green

Fig. 2

Effects of bias correction in overlapping regions of PE reads (schematic). Me: Methylated site. a The overlapping region of a pair of PE reads. The C bases in position a and c are not in overlapping regions. However, the C base in position b is in an overlapping region of the pair of reads. b A case assuming methylated reads have the same proportion as unmethylated reads (i.e., methylation rate is 50 %). c The upper panel shows a case assuming that unmethylated samples were sequenced by PE sequencing and methylated samples were sequenced by SE sequencing. The lower panel shows a case in which unmethylated samples were sequenced by SE sequencing whereas methylated samples were sequenced by PE sequencing

Fig. 3

ASM detection. Purple Cs are methylated, whereas red Ts are not methylated. Me: Methylated site. a Basic case in which two C bases are methylated. b An example of an ASM region in a monoclonal tumor marked by a heterozygous G/T SNP. c An example of a polyclonal tumor in which the heterozygous SNP and reference allele are present. d An example of another type of polyclonal tumor in which the heterozygous SNP was changed to a homozygous G allele due to loss of heterozygosity