Since the validity of a color-space mismatch depends on whether neighboring colors have been attributed to BINS, SOCS-B uses a dynamic programming table to compute the most probable methylation state for each cytosine based on the quality scores for each color (Supplementary Fig. S1).
The variant mode "short variants" was set for detecting valid sequence-space mismatches indicated by color-space mismatches.
In color-space, a nucleotide substitution appears as a specific pattern of two adjacent color-space mismatches.
2012) [ 44] with the following parameters: three color-space mismatches within the 'seed sequence' (first 18 bases of the reads), and six color-space mismatches on the following positions of the 35 bp reads.
The alignment of the experimental 50 nt-reads was set to tolerate a maximum of six color-space mismatches.
The results were then outputted as the coverage at each position and the valid sequence-space mismatches detected from each strand of the chromosome.
A total of 12,577,350 valid sequence-space mismatches (i.e., potential single-base substitutions) were detected out of 14,574,269,233 aligned bases.
The problem with this approach is that reads containing both methylated and unmethylated cytosines could contain numerous color-space mismatches against either reference sequence.
Sequence analysis was performed using the Small RNA Analysis Tool (RNA2MAP) [ 25] using the following parameters: three color-space mismatches within the 'seed sequence' (first 18 bases of the reads), and six color-space mismatches on the following positions of the 35 bp reads.
During alignment, up to 2 mismatches were allowed, treating valid-adjacent mismatches (those colour-space mismatches when located side-by side, indicate the presence of a single nucleotide variant) as a single mismatch.
However, in BS-Seq, bisulfite-induced nucleotide substitutions (BINS) are ubiquitous, causing most reads to contain too many color-space mismatches relative to the reference sequence to be aligned using the standard color-space alignment tools.
