The short answer is unfortunately "it can't be done". Here are at least two reasons why it cannot be done in the general case:
Orthology is not transitive. It is fully possible to have situations in which X is an ortholog of Y, Y is an ortholog of Z, but X is not an ortholog of Z. This effectively prevents making the type of table that you request with multiple species in different columns.
Orthology is not always one-to-one. It is not necessarily so that X in species 1 has only a single ortholog Y in species 2. One-to-many or many-to-many orthology can occur whenever gene duplications have taken place in one or both lineages since their divergence. This too will not fit with the table structure that you request.
It is also for these reasons that you cannot in general take the pairwise orthology assignments from e.g. Inparanoid and turn them into a single multi-species table.
You may in some special cases be able to get away with it, namely if you are looking at closely related species or highly conserved genes, in which case there will be so few gene duplications that orthology will in practice be one-to-one across all the species for almost all the genes. However, even in these cases you need to bear in mind that there will be exceptions and that you will have to deal with them.
agreed. From an evolutionary perspective, orthology is a very strong statement. It's associated with speciation event, which itself is a very complicated process.
Lars, don't you mean Homology is not transitive and not one-to-one. Ortholog is a very specific term indicating direct vertical transmission post speciation, that in itself guarantees a one-to-one correspondance. Anything else is a paralog, whether it was duplicated before or after the speciation event.
No, I actually do mean orthology, and it does not in any way guarantee one-to-one correspondance. Not if gene duplications take place after the speciation event, in which case you would get paralogous genes in one species that are all orthologs of the same gene in the other species.
Come to think of it, homology is transitive (as opposed to orthology). If A is homologous to B and B is homologous to C, I cannot think of a way that A would not be homologous to C.
Come to think of it, homology is transitive (as opposed to orthology). If A is homologous to B and B is homologous to C, I cannot think of a way that A would not be homologous to C. Unless you start looking at fusion proteins in which case different parts of the protein may have different evolutionary pasts.
After the speciation event is the key. Any new copy would have an independent evolutionary history from the moment of duplication (i.e. non-vertical transmission/horizontal transmission), since it simply did not exist before then. This would make it a paralog. You're absolutely right about homology being transitive, btw.
I'm sorry Paul, but we are not going to agree on this. The definition is very clear in Fitch, Systematic Zoology, 1970. See for example Figure 1 in http://www.ncbi.nlm.nih.gov/pmc/articles/PMC138949/ for a clear examples of orthology that is not 1:1.
I'm sorry Paul, but we are not going to agree on this. The definition is very clear in Fitch, Systematic Zoology, 1970. See for example Figure 1 in ncbi.nlm.nih.gov/pmc/articles/PMC138949 for a clear example of orthology that is not 1:1.
agreed. From an evolutionary perspective, orthology is a very strong statement. It's associated with speciation event, which itself is a very complicated process.
Lars, don't you mean Homology is not transitive and not one-to-one. Ortholog is a very specific term indicating direct vertical transmission post speciation, that in itself guarantees a one-to-one correspondance. Anything else is a paralog, whether it was duplicated before or after the speciation event.
No, I actually do mean orthology, and it does not in any way guarantee one-to-one correspondance. Not if gene duplications take place after the speciation event, in which case you would get paralogous genes in one species that are all orthologs of the same gene in the other species.
Come to think of it, homology is transitive (as opposed to orthology). If A is homologous to B and B is homologous to C, I cannot think of a way that A would not be homologous to C.
Come to think of it, homology is transitive (as opposed to orthology). If A is homologous to B and B is homologous to C, I cannot think of a way that A would not be homologous to C. Unless you start looking at fusion proteins in which case different parts of the protein may have different evolutionary pasts.
After the speciation event is the key. Any new copy would have an independent evolutionary history from the moment of duplication (i.e. non-vertical transmission/horizontal transmission), since it simply did not exist before then. This would make it a paralog. You're absolutely right about homology being transitive, btw.
I'm sorry Paul, but we are not going to agree on this. The definition is very clear in Fitch, Systematic Zoology, 1970. See for example Figure 1 in http://www.ncbi.nlm.nih.gov/pmc/articles/PMC138949/ for a clear examples of orthology that is not 1:1.
I'm sorry Paul, but we are not going to agree on this. The definition is very clear in Fitch, Systematic Zoology, 1970. See for example Figure 1 in ncbi.nlm.nih.gov/pmc/articles/PMC138949 for a clear example of orthology that is not 1:1.