Strictly speaking, you can never determine homology, with certainty, simply by comparing sequences over any evolutionary distance. Based on sequence similarity, we infer homology, but cannot prove it merely from any measure of sequence similarity. Also keeping in mind that homologous sequences can be orthologous (divergent from a common ancestor) or paralogous (a result of gene duplication) as well as other mechanisms or variations of those. And raw similarity can result from convergence as well. So trying to tease apart how a particular set of similar sequences actually came to be similar is one of those questions at the very heart of molecular phylogenetics. You are trying to make statements about past events for which you can never actually determine the true course of evolutionary history.

So you can attempt to test your own sequence's homology by various methods. Randomize the sequences and compare alignments of those to the original alignments (original should differ significantly from randomized sequence alignments). Compare phylogenetic trees inferred from your experimental sequences with independently inferred gene tree or species trees for the species from which you collected your samples. Consistent congruence of trees with independently inferred species trees would provide evidence for homology. Translate your gene sequences and compare the protein sequences and their alignment(s), and their inferred trees.

If you can do so, try to collect sequences for multiple genes, so you can compare or contrast the inferred trees for each to look for inconsistencies amongst them.

Try to include rational outliers in your species or taxa sampled so you can root your trees with realistic ancestral sequences (or use published sequences if they are available).

Ultimately though, to prove homology, you would have to know the underlying evolutionary history with certainty, which of course you never can. So all you can do is infer it and try to test your inference rationally in ways that would be expected to show inconsistencies if the sequences were not in fact homologous.

What I understand in simpler terms is, you want to check experimentally why this gene is conserved throughout evolution ? What factors made it conserved among a wide range of organisms ?

Kind of. I am going to be using genome wide sequencing to identify loci responsible for specific colour polymorphisms. What I want is a way to determine if I find such loci, say one responsible for banding, is it homologous with the Agouti gene seen in mammals or not. Is it the same genetic region, possessed through a shared common ancestor, or an independent evolution to the same phenotype. Likewise if I find a loci responsible for melanism is there a way to test if it is homologous MC1R in mammals or again unrelated, independent evolutions? Is there a test for this?

I hope that is clear, apologies if not.

I cannot tell for experimental validation, but if you had sequence in your hand, then you can easily play some bioinformatics with it to find its homologous sequences across the genomes, and when you got some, look for their loci and other details. Next you can compare all your homologs by Multiple Sequence Alignment, which will generate a phylogeny relation among the homologs. It will help you to find out which genes are close homologs and which ones are distant.

Strictly speaking, you can never determine homology, with certainty, simply by comparing sequences over any evolutionary distance. Based on sequence similarity, we infer homology, but cannot prove it merely from any measure of sequence similarity. Also keeping in mind that homologous sequences can be orthologous (divergent from a common ancestor) or paralogous (a result of gene duplication) as well as other mechanisms or variations of those. And raw similarity can result from convergence as well. So trying to tease apart how a particular set of similar sequences actually came to be similar is one of those questions at the very heart of molecular phylogenetics. You are trying to make statements about past events for which you can never actually determine the true course of evolutionary history.

So you can attempt to test your own sequence's homology by various methods. Randomize the sequences and compare alignments of those to the original alignments (original should differ significantly from randomized sequence alignments). Compare phylogenetic trees inferred from your experimental sequences with independently inferred gene tree or species trees for the species from which you collected your samples. Consistent congruence of trees with independently inferred species trees would provide evidence for homology. Translate your gene sequences and compare the protein sequences and their alignment(s), and their inferred trees.

If you can do so, try to collect sequences for multiple genes, so you can compare or contrast the inferred trees for each to look for inconsistencies amongst them.

Try to include rational outliers in your species or taxa sampled so you can root your trees with realistic ancestral sequences (or use published sequences if they are available).

Ultimately though, to prove homology, you would have to know the underlying evolutionary history with certainty, which of course you never can. So all you can do is infer it and try to test your inference rationally in ways that would be expected to show inconsistencies if the sequences were not in fact homologous.