Biology Reference
In-Depth Information
•
average distance between internal loops;
•
average size of symmetrical loops;
•
average size of asymmetrical loops;
•
proportion of A/C/G/U nucleotides in the stem; and
•
proportion of A-U/C-G/G-U base pairs in the stem.
A second subset captured information about the longest symmetrical
region in the stem, i.e. the longest region devoid of bulges or asymmet-
rical loops, and information about the longest region with relaxed sym-
metry, i.e. the region in which the difference in the number of unpaired
nucleotides on the 5
′
and the 3
′
arm of the stem was below a given
threshold:
•
length;
•
distance from the hairpin loop;
•
number of nucleotides involved in internal loops (calculated sepa-
rately for symmetrical and asymmetrical loops in the case of the
region with relaxed symmetry);
•
proportion of A/C/G/U nucleotides; and
•
proportion of A-U/C-G/G-U base pairs.
Finally, another subset of features captured information determined
from sliding windows of the length of the mature miRNA (22 nucleotides):
•
maximum number of base pairs;
•
minimum number of nucleotides in asymmetrical loops; and
•
minimum asymmetry over the internal loops in this region.
The model obtained this way was used to score robust stems
extracted from a large panel of animal pathogenic viruses. miRNAs were
predicted in viruses of the herpes and polyoma families, with overall sen-
sitivities and specificities around 50%.
49
Finally, a fundamentally different approach to miRNA gene prediction,
moving away from biogenesis requirements and focusing on the effector
arm of the miRNA pathway, was taken by Xie
et al
.
101
Since various stud-
ies indicated that seven to eight nucleotides at the 5
′
end of the miRNA
