Is it possible to design universal primers for detection of Pseudomonas spp.?

Question

[Alec Watanabe]

Hi,

A professor asked me to design primers for the detection of Pseudomonas spp.. The idea is to detect Pseudomonas reads in cancer samples. According to an article, many bacteria could be potentially involved in cancer besides H. pylori. I am not sure if designing primers will assure the detection of the bacteria since we would with reads. If the objective is to detect Pseudomonas spp. in cancer samples, the bacteria reads from one sample will probably differ from another sample. I am not an expert in primer design so what is your opinion on this topic?

Let's say it is possible to design primers for detection of Pseudomonas spp. in different cancer samples, what would be the in silico approach to accomplish the primer design? I initially thought about analyzing the whole genome sequences of Pseudomonas spp. but the analysis is extremely extensive (for both the average computers and humans). The second strategy I came up with is to analyze specific gene sequences like 16S rRNA. This strategy would analyze several 16S rRNA sequences from the genus to find conserved regions and design primers. Unfortunately, after a quick analysis using MUSCLE for sequence alignment and UGENE for alignment visualization, there are no truly identical sequences between the 288 Pseudomonas sequences I analyzed. Even in "conserved regions", one sequence is different than the others or the conserved region does not fit the ideal primer characteristics (e.g. Tm, length, GC content, less than 4 consecutive G residues, etc). Are universal primers truly universal? What's your ideas on designing universal primers for the identification of bacteria at the genus level?

[GenoMax]

Detection of pathogens using qPCR has been around for a while. If you do a google search with improved pcr for identification of pseudomonas aeruginosa you will find plenty of papers.

Here is a review on real time PCR for microbial detection that you will find useful.

This thread follows-up from a previous discussion : Is it possible to use local blastn to obtain the aligned sequences of complete record files?

[microfuge]

Not a direct answer but NCBI Probe dataset also provides primer sequences For Pseudomonas ( " https://www.ncbi.nlm.nih.gov/probe/?term=txid286[Organism:exp] " ) . I would suggest to start with NCBI taxonomy to get to the species/genus node and then use the linkout to probe dataset to look into studies which involves the node.

Answer 1

[Mensur Dlakic]

I would start by finding ultraconserved elements in all Pseudomonas spp. Particularly, I would look into intergenic spacers (non-coding parts of the genome; promoters). Next I would compare the sequences to all bacterial genomes and select those that match only Pseudomonas spp. I am assuming you know how to design primers once you have the sequence(s).

[Alec Watanabe]

Hi Mensur,

Assuming we would search for ultraconserved elements in all Pseudomonas spp., would you run a global alignment first? If so, would an average computer have enough computation power to handle the high number of genomes? I am familiar with the subsequent steps you mentioned.

Thank you for your constructive critic. The post is now edited.

[Mensur Dlakic]

I don't think you need a global alignment. As I understand it, your goal is to find a relatively short and unique piece of DNA shared by multiple genomes. That can be done by looking either at genes or intragenic regions.

That ultraconserved elements has links to programs that you may want to check out. It could automate some steps and make it faster than a manual procedure I will describe below.

Old-style NCBI annotations (see here) used to have a set of all genes in .ffn files. Not sure that more recent annotations have it. I have custom scripts and programs that extract genes and intergenic regions from genomes, but they are too clunky to share. There should be tools for that available as this is a very common research problem. If all else fails, there is a large set of all Pseudomonas genes and intergenic regions at pseudomonas.com. Google will help you find tools to convert .gff files to .fasta format.

Assuming you have all intergenic regions, drop all sequences shorter than say 100bp, make a blast-readable database and BLASTn each of these regions against it. The goal is to find those intergenic regions that can be found in all Pseudomonas genomes and share high identity - I'd look for >80% identity, a length of at least 100-150bp, and of course that they are not tRNAs or rRNAs. I did this more than a decade ago on 7 or 8 Pseudomonas species and there were many regions that fit the bill. Not sure if the same will be true for hundreds of species/strains, but I suspect you'll still end up with a decent number of candidates.

As I said before, the last step would be to take these sequences that are shared among all or most Pseudomonads, BLASTn them against the NT database, and select those regions that have no matches outside of Pseudomonads.

Everything that I described above can be done with a set of genes rather than intergenic regions. I think intergenic regions are better first choice because it is less likely that a highly conserved gene will not be found in any other species.