Biomedical Engineering Reference
In-Depth Information
accessibility within the 3 UTR was further underlined by the observation that these
functional target sites seem to be under evolutionary pressure (Kertesz et al. 2007 ).
A large proportion of target algorithms, therefore, continue to place emphasis on
the 3 end of the transcript during scoring. However these sites are not solely confined
to the 3 UTR and recent evidence points to biologically active sites located within
both the 5 UTR and coding sequence (CDS) (Moretti et al. 2010 ; Schnall-Levin et al.
2011 ). Sites within these regions seem to occur less frequently and are on the whole
less effective than those within 3 UTR (Bartel 2009 ), although target binding sites
within the 5 UTR of the transcript can in some cases disrupt translation as efficiently
as those within the 3 UTR (Lytle et al. 2007 ) and in some instances simultaneous
targeting of both the 5 and 3 UTR by a single miRNA can also occur (Lee et al.
2009 ).
3.2.3
Cross Species Conservation
The number of common miRNA targets predicted by the early algorithms following
analysis of the same species was low. To alleviate these concerns a further step was
introduced to determine evolutionary conservation across multiple species which
allowed targets that scored equally well in sequence alignment to be ranked according
to conservation (Bartel 2009 ). The rationale behind the use of conservation analysis
of transcripts and miRNA target regions is that miRNA-mRNA interactions that
are present across a large evolutionary distance suggests biological functionality.
The majority of target algorithms utilise a combination of seed region pairing and
phylogenetic filtering to reduce false positive results. While conservation provides a
strong indication for functional relevance, not all functional targets of miRNAs are
conserved (Farh et al. 2005 ).
3.2.4
Thermodynamic Calculations
In recent years significant false positive rates associated with seed rule/conservation
based approaches have driven the development of algorithms that assess the binding
efficiency of a miRNA with its target. Following the observation that the free energy of
hybridisation between the miRNA and mRNA target site ( G duplex ) was an important
factor in determining biological function (Doench and Sharp, 2004 ), thermodynamic
calculations were incorporated into several algorithms. G duplex can be calculated for
each predicted miRNA-mRNA pair, the most commonly used program for this task is
known as the Vienna package (http://www.tbi.univie.ac.at/RNA/) (Hofacker 2003 ).
The lower the free energy the greater the likelihood of an interaction; a G duplex
threshold can be applied to remove putative target interactions above a free energy
level. miRNA target prediction schemes tend to vary this threshold according to
species, e.g. Drosophila —14 kcal/mol (Enright et al. 2003 ), Human—17 kcal/mol
Search WWH ::




Custom Search