Find correlated genes with a metabolite across different treatments

Question

[khhgng]

Hi.

I have a situation where a sample material is given 4-5 types of treatment and the transcriptomic data has been obtained. All of the treatments induce certain set of genes (redundant between treatments) that may lead to production of certain metabolites or say amino acids.

How can I use this set of genes to identify other genes in all these treated samples that may be correlated with this set of genes ? Can WGCNA be used for this ?

Further how can I use these set of correlated genes to know if they are co-regulated also ?

Thank you

[cpad0112]

All of the treatments induce certain set of genes (redundant between treatments) that may lead to production of certain metabolites or say amino acids.

From your transcriptomic data, were you able to confirm this?

If it reflects your study goal, then look at the differential expression patterns of all those genes involved in pathways/networks of certain metabolites / aminoacids.

[khhgng]

Yes, at least the analysis of metabolic pathways in the transcriptomic data highlights this. Though not the same set of genes is highlighted between treatments but they all map to biosynthetic pathway of this specific metabolite.

Answer 1

[theobroma22]

Have you heard of clustering your genes based on expression patterns? If they cluster together they are correlated or co-regulated. You can then take each cluster and characterize the contents using gene ontology or other. Alternatively, if you have the Entrez IDs I recommend for you to use the SPIA package available on Bioconductor.

[khhgng]

Thanks. In this case each treatment has 4000-6000 genes DE (with 3 replicates). Can you recommend what method or programme can be useful for such clustering ?

Is there an easy way out to follow a subset of genes redundant in expression that are associated with certain metabolite biosynthesis ? I am looking for genes not just involved in the metabolic pathway but also other genes that may be associated with that in all these treatments when this metabolite is synthesized.

[theobroma22]

You can use the Mfuzz package available on Bioconductor to cluster your data. A way out would be to use RTPCR of the genes you identified in your treatments producing that metabolite / AA, since you must validate your transcriptome data for publication. A challenge is to make sure you have the right gene isoforms among the different treatments for each pathway you are investigating.

[khhgng]

Thanks. I'll take the gene ontology route first and try to look for genes based on GO.