NCBI Primer Design - Stringency Issues - (Feb/26/2014 )

Hi All,

 

I'm using PrimerDesign to create primer pairs specific to a region of the x-chromosome of canis familiaris.  I'm having some issues with stringency/specificity.  I believe I am selecting for the most stringent primers,  yet, if I run the generated primers through PrimerDesign a second time, multiple hits (sometimes 20+) with only a few mismatches are found.  Here are the parameters I'm using, if not specified, it was left as default:

 

PCR Template:

Refseq Record: NC_006621.3 (Canis Lupus Familiaris X-Chromosome Refseq)

Forward Primer Range: 31869000-31869499

Reverse Primer Range: 31869500-31869999

 

Primer Parameters:

PCR Product Size: 350-1000

# Primers to Return: 10

 

Primer Pair Specificity Checking Parameters:

Specificity Check: On

Database: Genome (chromosomes from all organisms) - using refseq results in "Empty Result" error

Organism: Canis lupus familiaris (taxid:9615)

Primer specificity stringency: Primer must have 6 total mismatches to unintended targets, including at least 5 mismatches within the last 5 bps at the 3' end.

Ignore targets that have 9 or more mismatches to the primer.

 

Advanced Parameters

 

Primer Pair Specificity Checking Parameters

Max number of Blast target sequences: 100,000

Blast expect (E) value: 100,000

Blast word size: 6

Max primer pairs to screen:  2,000

 

With these values, PrimerDesign returns 10 pairs, the first of which is:

 

Forward: GTTCTCTTTGTGGCTGTGGC

Reverse: CTCCGAGCTTGGACACCAAC

Product length: 709

 

If you then take this primer pair, and enter a new PrimerDesign window with the following conditions:

 

Primer Parameters

Use my own forward primer: GTTCTCTTTGTGGCTGTGGC

Use my own reverse primer: CTCCGAGCTTGGACACCAAC

 

Primer Pair Specificity Checking Parameters

Specificity Check: On

Database: Genome (Chromosomes from all organisms)

Organism: canis lupus familiaris (taxid: 9615)

 

The results are frustrating.  There are over 20 non-specific amplicons listed.  They vary in size 400-4k+.   They vary in mismatches - anywhere from 0-6.  And there are not 5+ mismatches at the 3' end of the primers.  Right now I'm forced to double check each of the 10 primer pairs, and hope that one of them is within the specificity range I've requested.  It's time consuming - please, tell me there's a better way.

 

Can someone please explain the parameters I should be using to design specific primers.  I'm amplifying from whole dog genome DNA, so specificity is imperative (I'm also in what seems to be a repetitive region).

 

I don't think those stringency parameters are to be taken as an unbreakable rule.

 

There are many things that matter. The length of the unspecific product, if its very close to your, it's a problem, since you may not see unspecifities on the gel. If it's more than 2000 it usually doesn't matter, if you use conditions that suits your product length.

Shorter product than intended generally also need attention, since they would preferentially amplify.

 

I have experience, that 2 mismatches near the end are quite enough to not amplify.. but the aforementioned things have to be taken into account.

 

Also you need take the binding of both primers into account. If one is almost specific and the other one has many mismatches, the product is unlikely to be created.

 

I'm not sure what PrimerDesign is exactly, but obviously it seems to run all the data through PrimerBLAST.

 

I put your parameters there (you can try filling the Job ID : MyjeV5VdPHkDVAdYZnY1Jn1bBzduRBoy to see yourself), and the primers you mentioned were as a 3rd primer pair. There are many >2kb products I will not even consider, but also 4 smaller, some of which are not quite safe AFAIK.

 

The primer designed as a first one has better parameters, and less considerable unspecifities.

But, If you choose to check those selected, BLAST will also give you a different result. I suppose it's because it doesn't show all but somehow desiced, which to display, and displays different each time. I think it's a feature of BLAST (which is shared in your design program, which uses it). If I limit unspecifities only to 2kb I get only three unspecific product, smaller than target, which have either mismatch right at 3' end (plus some further to 5') or more mismatches in the middle, and I think using a decently stringent PCR conditions, it would be possible to amplify the specific product.

 

Seems there is not much better way to find specific primers by automatic design.

Other way would be to look through the similar targets, that worries you most and analyze both regions (specific and nonspecific, where it would be most suitable to design specific primer, bit this is really laborious task, and I only do that in cases where there is a grave similarity to my target sequence)

Trof on Thu Feb 27 13:25:20 2014 said:

I don't think those stringency parameters are to be taken as an unbreakable rule.

 

There are many things that matter. The length of the unspecific product, if its very close to your, it's a problem, since you may not see unspecifities on the gel. If it's more than 2000 it usually doesn't matter, if you use conditions that suits your product length.

Shorter product than intended generally also need attention, since they would preferentially amplify.

 

I have experience, that 2 mismatches near the end are quite enough to not amplify.. but the aforementioned things have to be taken into account.

 

Also you need take the binding of both primers into account. If one is almost specific and the other one has many mismatches, the product is unlikely to be created.

 

I'm not sure what PrimerDesign is exactly, but obviously it seems to run all the data through PrimerBLAST.

 

I put your parameters there (you can try filling the Job ID : MyjeV5VdPHkDVAdYZnY1Jn1bBzduRBoy to see yourself), and the primers you mentioned were as a 3rd primer pair. There are many >2kb products I will not even consider, but also 4 smaller, some of which are not quite safe AFAIK.

 

The primer designed as a first one has better parameters, and less considerable unspecifities.

But, If you choose to check those selected, BLAST will also give you a different result. I suppose it's because it doesn't show all but somehow desiced, which to display, and displays different each time. I think it's a feature of BLAST (which is shared in your design program, which uses it). If I limit unspecifities only to 2kb I get only three unspecific product, smaller than target, which have either mismatch right at 3' end (plus some further to 5') or more mismatches in the middle, and I think using a decently stringent PCR conditions, it would be possible to amplify the specific product.

 

Seems there is not much better way to find specific primers by automatic design.

Other way would be to look through the similar targets, that worries you most and analyze both regions (specific and nonspecific, where it would be most suitable to design specific primer, bit this is really laborious task, and I only do that in cases where there is a grave similarity to my target sequence)

Hi trof, thanks for your response.  You're right that products not near my amplicon size can be ignored.  However, I'd like to send some of these for sequencing in the future.  The product size will be unknown (long story, working with a deletion).  I suppose an option would be to do multiple runs, with a new forward/reverse primer, that moves inward each round, and then use yet another primer for sanger sequencing.

If you sequence only a gel-excised specific fragment, your sequencing primers need to be even less specific. There is no other template to bind.