Dear Shilpi Nagpal,

Generally, we use jModel test to find out the best fit models. It may require when you are going to conduct some phylogenetic analyses. In many cases, it may need to mention the best fit models before running the phylogenetic analyses (eg., BI analyses). So far, there is nothing to do with this after preparing the phylogenetic tree. However, for RAxML, if you are dealing with a single gene, you may leave most of the options as default. You will just need to convert your aligned fasta file to phylip file, upload it, select bootstraping into fixed and submit your data. Hope this answer will help you.

Best regards-

Yes. Prior to creating a phylogenetic tree, you have to determine the best-fit model, which is estimated based on the number of sequences involved, arrangement/alignment of bases and the nucleotides composition/length of your multiple sequence alignment file. Such best-fit model will be applied in a phylogenetic tree software to reconstruct the relationship between the species included in your alignment based on several statistic parameters. J-modeltest is one such software. The result of a J-modeltest analysis is a list of best-fit models, sorted from the best to the least-fit models. Sometimes, reviewer of a journak will ask about this model inference and its statistical parameters.

However, if you use RAxML, there is only one type of best-fit model that can be applied to reconstruct a phylogenetic tree from your alignment file, which is the GTR models (e.g. GTR, GTR+I, GTR+G, GTR+I+G). The problem is, J-Modeltest may suggest a best-fit model different from GTR type (e.g. JC, HKY, TrN, etc.), thus such model can not be used in RAxML. Alternatively, you can use PhyML to be able to use different best-fit model for your phylogenetic reconstruction. Or just use MEGA X which is a one-for-all software for phylogenetic analysis, therefore is more convenient. Cheers.