Dear Neha,
I would first check the functional domains of your protein. You can do this from Uniprot website, InterPro, etc.
Afterwards I would check the nature of the mutation. Is this a frameshift mutation somewhere in the middle of the protein? Then it is more likely that it will result in a premature stop codon. Is it a single nucleotide variation? Then you should see if this mutation is located in one of the functional domains above.
If it is located in the functional domains, than check if mutations in similar proteins in the similar region also result in similar phenotype. Next step is to examine the region where the mutation occurs. What type of amino acids are there? Is it enriched in polar, non-polar, negatively charged amino acids.... etc. And compare this to the new amino-acid. Does it replace an amino acid that had different chemical features etc.
Next is to see and predict whether the mutation results in a change in 3D structure of the protein.And this you can do with freely available online 3D modellers. Like I-TASSER, etc. Compare your 3D structures with 3D structures of similar proteins bearing mutations in those functional domains. Then see if the changes can tell you something taking the 3D model of the wild-type as reference. You can also compare the secondary structures of resulting models. Sometimes they tell a clearer story.
When it comes to prediction, the list never actually ends, and there is no optimal strategy. But for now, such in-silico approach is as good as it can get.
I hope this helps,
Good luck with your research