Sequence Analyses For Microarray Design

Question

[Nina]

I'd like to know the feasibility of designing an array for multiple species (5-6) of the same genus, for expression analysis. There are only 5 species with fully sequenced genomes and a lot of contings available for this genus. (I'm not able to use Perl). My questions are: - it is possible with my low bionformatic knowledge about informatic languages to manage and design an array for about 300 genes for 5-6 different species? - I know the possibility to build a local blast database and work with open source oligoarray design, but how can I select conserved sequence for each gene among the 5 species? I have to use clustalW each time to analyse each gene? There are softwares or algorithms that can do it? - Probably could be better to spot a different oligo for each gene and each species to be sure to do not have false negatives?

Answer 1

[Neilfws]

To be frank, without some bioinformatics knowledge, you will find this task rather difficult. There are, for example, many better and more appropriate multiple alignment tools than clustalW.

However, since array design is a common task, there will certainly be software available to help you do it, which may or may not be free. I'd start simply with a Google search for microarray design software, see what's out there and whether any of it is appropriate for you. Some results to get you started:

• Primer design software
• Primer and oligo design tools
• SMD software and tools

If that doesn't get you anywhere, seek out a knowledgeable expert (locally or online) who can help. It's often more productive to strike up a good working relationship with someone who knows the area well than to try and do everything yourself.

Answer 2

[biobot 0.0.77.a.1099]

You should seek expert help for an additional reason; the analysis strategy for small, specialised (often called 'boutique') arrays will need to be very different from that used for say, Affymetrix, Illumina or large spotted gene expression arrays. When the intensities of the majority of the probes on the array are changing between experimental conditions, the assumptions made by some normalisation methods are violated and you will get misleading results if you use them. See this paper from Gordon Smyth's group.

So, not only do you need to consider oligo design, but also your proposed method of analysis.

1. Plan your analysis before you design your array

2. Plan your experimental design and replication strategy before you obtain any biological samples