How can I find out if my gene of interest is duplicated?

Question

To identify homologous genes, I performed tBLASTn analyses against my genome of interest resulting in sequences of a closely related species. Using this method, I found two (and in some cases up to five copies) of my different genes of interest in the target genome.

How can I determine if these gene copies correspond to true duplication events?

I aligned the newly found sequences to the reference genes and I found good alignments in more than one region (20% - 50% identity), but I am unsure if this is enough to consider them as duplicates.

Is there a threshold to help me determine if they are indeed duplications? How can I distinguish between a true duplication event and a pseudogene or a paralogous gene? The target genome belongs to a polyploid organism, although the genome assembly type is listed as haploid.

Any help will be greatly appreciated.

---

Thank you all very much for your responses, I am now beginning to understand how to find out if my genes are duplicated.

I performed some of the analyses that you suggested and I would very much appreciate your input on the results.

First, I followed @terdon's advice and I took the best hit for one of the query (source) genes in the target genome and I used this sequence as a query against the query (source) genome, i.e., I performed a reciprocal BLAST analysis (in my case it was a reciprocal BLASTp analysis). The first hit was my original query gene, which had an E-value of 0.0. For other sequences the reciprocal BLASTp hits were more variable, ranging from 3e-42 to 2e-154, but the first hit was always the query sequence. I also looked at the protein domains and they are conserved, i.e., both the query and the target sequences have the same domains. As for the exonic structure, it is fairly similar, for e.g., the query has 12 exons while the target sequence has 10. Exons are similar in length. So, in summary, when the potential duplicates from my target genome are blasted against the original source genome, the best hit is the original query sequence.

As for @zhanf's suggestion, I did a basic phylogenetic analysis using Clustal Omega (neighbour-joining tree without distance corrections) using four sequences: the original query sequence and the three gene copies (seq_5-1, seq_5-2 and seq_5-3) that I found in the target genome. This is the result:

I also compared each of the sequences against the original query:

• seq_5-1 vs query: 34% identity (query length 780 aa, seq_5-1 length 400 aa)
• seq_5-2 vs query: 42% identity (query length 780 aa, seq_5-2 length 580 aa)
• seq_5-3 vs query: 40% identity (query length 780 aa, seq_5-3 length 490 aa)

So in summary, these sequences would appear to be paralogues, right?

Again, thank you very much for your help, I greatly appreciate it!

Leah

Answer 1

You might want to have a look at What is the difference between orthologs, paralogs and homologs?. Paralogs are duplicates by definition, and whether a gene is a pseudogene or still active has no bearing on whether it comes from a duplication.

In any case, to figure out what's going in in your target genome, you should use the target's genes. So, first get the best hit for each of your query genes in the target genome. Then, to investigate its evolutionary history, use that sequence (the one from the target genome) as query against the target genome. That will let you assess how similar the target's hits are to one another and that's how you can attempt to figure out their evolutionary history.

Note that 20% identity is very low for closely related species. While there is no golden threshold, I would expect recent duplication events (and in closely related species, a duplication event that is in one genome and not the other will be recent) to produce much more similar sequences. In any case, you can't do this based on sequence identity alone.

Are the protein domains of the gene products conserved? Is the exonic structure of the gene conserved? You should also try taking the putative duplicate of geneA from genome2 and blast it against genome1 (your original source genome). Is the best hit the geneA from genome1, the query you used to find the putative duplicate in genome2? Or is there a better hit for it against another gene?

These are all questions to investigate and take into account when reconstructing the evolutionary history. You can't just look at identity and use some threshold to decide this sort of thing.

If you blast the potential dupicate from your target genome against your original source genome, is the best hit the duplicate?

Answer 2

I don't have enough reputation to comment but if a sequence is duplicated then there has obviously been a duplication event. Doesn't matter how scientists call it now.It can be older or more recent but it's yet a duplication.