Biology Reference
In-Depth Information
UPGMA to construct a guided tree and a progressive alignment is
performed by considering only subtrees. Finally, in the refinement
step, an edge from the previous guided tree is deleted. This splits
the guided tree into two subtrees and the profile of the multiple
alignment for each subtree is calculated. The new MSA produced
by realigning the two profiles is compared with the previous align-
ment. If the SP score of the new MSA is improved, then alignment
of the two profiles is kept for the next iteration. Refinement phase is
repeated until convergence is achieved.
The term profile alignment resembles alignment of two align-
ments by matching up corresponding columns and gives scores
based on composition of columns. Profile sum of pairs function
and log-expectation scoring function of MUSCLE are defined as
follows:
X
X
X
X
p
ij
p
i
p
j
PSP
XY
f
i
X
f
j
Y
S
ij
¼
f
i
X
f
j
Y
log
¼
i
j
i
j
log
X
i
X
p
ij
p
i
p
j
LE
XY
f
G
Þð
f
G
Þ
f
X
f
Y
¼ð
1
1
i
i
j
where
f
G
X
is frequency of gaps in profile column
X
.
4 Conclusions
Evolutionary and functional relationship between sequences can
be inferred by assessing their alignments. However, finding the
optimal MSA between a set of sequences is an NP-complete
problem and assessing the results of a multiple sequence align-
ment is not an easy task as in the pairwise sequence alignment.
Determining the performance and accuracy of an MSA relies on
how diverged the given sequences are. Scoring functions give an
insight to assess the accuracy of an alignment and highlight vari-
ous criteria that a given algorithm tries to optimize. The MSA
algorithms used in practice employ heuristic solutions; therefore,
in most cases a manual inspection and intervention in parameter,
data set, scoring/objective function, and algorithm selection
would help enhance the overall performance. There exist MSA
editors such as JalView [
20
] that can be used to inspect and amend
alignment result. To best interpret the results of an MSA, an
expert opinion that factors in biological basis in the findings
constitutes a critical and valuable step.
