Defining New Gene Clusters

Question

[Rahul Sharma]

Hi,

I am interested in defining gene clusters(Gene having same functions/protein products) in a nematode genome. Is there any available software/methods defining these clusters? I have gone through some articles. They defined gene clusters, if 3 genes in a row having same function. Is there any possibility with sliding window approach? Lets say in a window of 5kb, if we get 3-4 genes (with same protein product) we can assign a cluster/island. Please suggest me other possible ways and any article related to this kind of cluster study.

Many thanks in advance, Rahul

Answer 1

[Alex Reynolds]

This paper on cell-cycle yeast genes suggests a method of hierarchical clustering by phased expression patterns. Once the clusters were calculated, the authors investigated and classified groups of cycle-related genes by function, such as metabolism or DNA replication, etc. Perhaps it might inspire further research on your part, based on your data.

Answer 2

[Josh Herr]

I'm not entirely clear on what you are asking to do. If you're trying to identify clusters of "genes having same functions/protein products" I would think you would have to show homology to known (or even uncharacterized, but homologous) motifs. I would do this using similarity searchers (BLAST, BLAT, etc.) and ultimately with phylogenetic methods. Both homologous search methods such as BLAST (that infers homology) and phylogenetic methods are covered with daily questions here in this forum.

[Rahul Sharma]

HI, thanks for your reply. By cluster I mean, I have few genes with known and similar function. Oftenly these genes are randomly distributed in the genome, but in some cases they are located very close to each other. The aim of the study is to define clusters of these closely located genes using some statistics. As I already mentioned, sliding window is a possibility. Is there any other way of doing this? Actually I used orthomcl/Inparanoid for ortho predictions, blast will not work in this case. Wishes, Rahul

[Josh Herr]

I think you'll still find the best way will be to ultimately use phylogenetic methods: you can see if your "closely located genes" are the result of tandem repeats or some selection factor on duplication events after a "gene cassette" has been translocated.

[Rahul Sharma]

Thank you so much for your message. Its interesting to look into the tandem repeats and duplication events.

Answer 3

[Alex Reynolds]

One reason that windowing might not give a complete answer is long-range regulatory interactions that would fall outside this window. The Dekker lab has investigated 5C-based approaches to this study. Their work may also be generally insightful.

[Rahul Sharma]

Thanks for your reply, Its very interesting.