An outgroup is a taxon that is more distantly related to the ingroup taxa than they are to each other.  While that sounds simple, there are some tricky aspects to choosing an outgroup.

The fist is that, when dealing with sequences, the outgroup taxon MUST be chosen on the basis of information that is EXTERNAL to the data set.  You can't simply choose a sequence that seems to be the most different from the others. If all your sequences (taxa) are from one species, picking a sequence from another species is a safe bet.  If all your sequences from the same family, you need to choose a sequence from a different family.  In both cases you are using well established taxonomic information to be confident that the ingroup taxa are more closely related to each other than they are to the chosen outgroup taxon.

Now comes the really tricky part: the outgroup taxon (or taxa) can't be so distantly related to the ingroup taxa that you can't get a reliable alignment.  That is, there must still be enough homologous characters (sites) to ensure that your tree is based upon comparison of homologous sites.  The more distantly related is your outgroup taxon the less reliable your alignment is likely to be.

For your bacterial 16S sequences you might turn to archaea for an outgroup

If you ingroup consists entirely of hominid sequences then you surely don't want to choose a snail sequence as your outgroup.  You want to pick a sequence from a species that is in the family most closely related to hominids.  (I have not idea what that would be.  I deal with bacteria)

Finally, you need to be sure that when you align the outgroup together with the ingroup sequences the alignment is reliable - or at least that there is a sufficient portion of the alignment that is reliable that you can use that part of the alignment to estimate your tree.

There is a web-based alignment program called Guidance that is very helpful with that aspect (http://guidance.tau.ac.il/ver2/).

Hi Neha,

In terms of outgoup you can use any species or species not similar for your sequences. You can use available sequences from GeneBank. Hope this clear?

Good luck 

Ali

Also you can look for this discussion linked below for more information.

https://www.researchgate.net/post/Which_would_be_the_best_outgroup_for_the_following_bacterial_strains

Good luck

Ali

For outgroup selection during phylogenetic analysis, we always use the most relative taxa of your target group. Thus it largely depend on your target group. Likewise, If your study model is a family A, then the best outgroup should be the member from the closet family of A.

Jie

hi, to all thank you for your valuable comments, here i am using 16srRNA bacterial sequences which include both gram positive as well as negative bacteria.. i took the whole phylum for phylogeny. so do you guys suggest that i must go for family phylogeny construction or genus level?? which method would be good for phylogeny as i am not sure for what analysis i should go.. kindly help..

It depends about what you have to phylogeny in my samples ( insects) I have used species level but you can start from genus level and see the results and that can help you to make your decision to go up for family or down to species.  Good luck 

Ali 

Hi Ali,

thank you so very much for your help. can you share your paper for reference..

An outgroup is a taxon that is more distantly related to the ingroup taxa than they are to each other.  While that sounds simple, there are some tricky aspects to choosing an outgroup.

The fist is that, when dealing with sequences, the outgroup taxon MUST be chosen on the basis of information that is EXTERNAL to the data set.  You can't simply choose a sequence that seems to be the most different from the others. If all your sequences (taxa) are from one species, picking a sequence from another species is a safe bet.  If all your sequences from the same family, you need to choose a sequence from a different family.  In both cases you are using well established taxonomic information to be confident that the ingroup taxa are more closely related to each other than they are to the chosen outgroup taxon.

Now comes the really tricky part: the outgroup taxon (or taxa) can't be so distantly related to the ingroup taxa that you can't get a reliable alignment.  That is, there must still be enough homologous characters (sites) to ensure that your tree is based upon comparison of homologous sites.  The more distantly related is your outgroup taxon the less reliable your alignment is likely to be.

For your bacterial 16S sequences you might turn to archaea for an outgroup

If you ingroup consists entirely of hominid sequences then you surely don't want to choose a snail sequence as your outgroup.  You want to pick a sequence from a species that is in the family most closely related to hominids.  (I have not idea what that would be.  I deal with bacteria)

Finally, you need to be sure that when you align the outgroup together with the ingroup sequences the alignment is reliable - or at least that there is a sufficient portion of the alignment that is reliable that you can use that part of the alignment to estimate your tree.

There is a web-based alignment program called Guidance that is very helpful with that aspect (http://guidance.tau.ac.il/ver2/).