Introduction
The PAR algorithm...this is where the magic happens in Bio::POD. You may have
dozens or even hundreds of genes that you are interested in screening. In addition,
you may also have hundreds of cases and controls to be screened as part of each
experiment. So when you choose a location to screen, that decision has a significant
amount of time and money attached to it. PAR can help by prioritizing where
to begin your search.
What is PAR?
PAR stands for Prioritization of Annotated Regions. It is an algorithm that
implicates regions of importance by the presence of known annotation such as
domains, secondary structures, SNPs, transmembrane domains, CG dinucleotides,
etc. To get a little more technical, it performs a summation over a pre-defined
"PAR window" of the gene, including annotation in the summation with
a score and weight for each annotation. The weight assigned to a type of annotation
affects how influential that type of annotation is on prioritizing your search.
There are two forms of PAR calculated in Bio::POD. Transcript PAR is calculated only for the exons of the gene and is displayed at the top of the Transcript View. Genomic PAR is calculated for the entire gene including introns and flanking sequence. The results of the PAR algorithm are displayed as graphs and are circled in the image below:
How do I interpret the PAR graph?
Essentially, larger PAR scores correlate with important regions of
the gene and should be considered for screening first. Therefore, the highest
peak on the PAR graph suggests the first region of the gene to begin your search.
This brings us to a discussion of "PAR regions". A PAR region is a sub-section of a gene that is suggested for screening. PAR regions have an order of importance and are displayed as vertical green column in Bio::POD. The 1st PAR region surrounds the highest peak of the PAR graph. The 2nd PAR region surrounds the next highest peak and so on. Two PAR regions are displayed and labeled for the Transcript PAR in the image above. The PAR regions can be interpreted as "Screen the 8th exon first, then screen the 9th exon."
Can I modify PAR?
Yes. PAR is meant to be customizable to your specific research interests. For
modifying PAR, you will want to be familiar with the "PAR Options..."
and "Annotation Options..." in the "Options" menu shown
here:
Let's start by clicking "PAR Options...". That will bring up the following dialog box:
The PAR Window Size is the width in nucleotides over which the summation is performed. You may wish to play with this number to see how it affects the PAR graph, but if you are unsure of what to use, simply use the default value of 150.
The rest of the options affect how the PAR regions are displayed. By default, they are displayed in "By regions" mode with a defined number of regions to be displayed and the maximum width defined in nucleotides. If you tend to span amplicon sizes of more or less that 150 nucleotides, you may which to change that number accordingly. The "By a threshold" mode will display a PAR region at every location in the graph that is above the specified threshold value.
Click "Ok" when you are satisfied with the PAR Options.
Next, click on "Annotation Options..." from the "Options" menu. The following dialog box will appear:
In this window, you may edit how each type of annotation affects the PAR calculations. Start by selecting the type of annotation from the drop-down menu at the top. In this example "domain" is selected.
The PAR-specific options are in the bottom half of the window starting with "Included in PAR calculations". If that box is checked, then this type of annotation will be included in the calculation. Alternatively, you may choose not to include this type of annotation in the algorithm.
By checking "Override score with", you may override the score for every annotation of this type (with a value between 0 and 1). For example, predicted secondary structure scores are determined by the prediction algorithm according to accuracy (they may predict accuracy of 70% resulting in a score of 0.7). You may override that score with this option.
And finally, you can change the default multiplier value. This determines the weight of the annotation type in the algorithm. If you think that domains highly correlate with the position of mutation for which you are searching, you may wish to increase the domain multiplier relative to the multiplier of other annotation types. You may enter a negative number if you think the mutation for which you are searching will be found in areas where this type of annotation is not located.
The Next Step
So what do you do with the PAR regions? The next step is to design primer pairs
for your experiment. You can select the sequence
in the PAR region, and span that selected region
with primers.