It uses a function M (s i 1, s i 2, j 1 ) which returns the index j2 of the character s i 2, j 2 of s i 2 aligned to s i 1, j 1 in OPA (s i 1, s i 2 ).
An alignment is characterized by the number of matches m, that of mismatches f, that of gaps g i in s i and that of gaps g j in s j.The length of s i is equal to m + f + g j and that of s j is m + f + g i, because any letter in s i is aligned to either a letter in s j or gap symbol in s j and vice versa.
After alignment, we filter (i) sequences aligned to more than one genomic location; (ii) sequences aligned to regions underrepresented in the reference sequence (ENCODE blacklisted regions); and (iii) duplicate reads to correct for PCR artifacts.
ts; and (ii) if F i is aligned to S/ S′, no F j for j> i is aligned to G/ G′.
The B score for the residue a i at position i in p is computed as follows: (1) where f (a i, h j ) is set to 1 if a i is aligned to an identical or similar residue in h j, or 0 otherwise, and c is a pseudo-count (set to 10 in this work).
Suppose, for example, that reaction R i appears in k reaction paths in R H 1 and that R i is aligned to R j k′ times (with k′≤ k) in the corresponding paths of R H 2 (through σ).
Finally, a constant gap penalty C f is applied to fusion transitions in which t i aligns to G/ G′ and t j aligns to S/ S′.
If the upstream end of node v i aligns to the downstream end of node v j, the edge in the graph would be e = (v i, v j ).
In other words, when symbol x i is aligned to a gap, we do not distinguish where is the location of this gap with respect to sequence Y. Columns (- j, i) are treated symmetrically.
A bed file containing 10,359 enriched TFII-I regions aligned to the human genome (hg19) was downloaded from GEO (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi acc=GSE51065) and input to ColoWeb.
