Trying to perform a branch-site analysis for positive selection
553 views
Skip to first unread message
Alexander McMenamin
Mar 4, 2020, 3:50:53 PM3/4/20
to PAML discussion group
Hi All,
I have a tree and an alignment which I successfully used for a standard analysis assuming one ratio (i.e., compares M0 and M2). But when I try to perform a branch-site model analysis all I get is an error saying Error: only two branch types are allowed for branch models". I can't find any info on what this might mean. Any help appreciated.
I have a tree and an alignment which I successfully used for a standard analysis assuming one ratio (i.e., compares M0 and M2). But when I try to perform a branch-site model analysis all I get is an error saying Error: only two branch types are allowed for branch models". I can't find any info on what this might mean. Any help appreciated.
Alex
Santiago Herrera
Mar 4, 2020, 4:00:24 PM3/4/20
to PAML discussion group
The branch-site model only allows to compare one "foreground" clade/branch against the rest of the tree ("backgorund" clades/branches) - where dn/ds is allowed to go >1 in the foreground branch - against a null model where codons are only allowed to evolve neutrally (dn/ds=1) or under purifying sln (0<dn/ds<1). My guess is that if you are specifying more than one foreground branch in the analysis that might be a problem.
Alexander McMenamin
Mar 4, 2020, 6:23:26 PM3/4/20
to PAML discussion group
Thank you, Santiago!
Ziheng
Apr 4, 2020, 11:42:45 AM4/4/20
to PAML discussion group
i am not sure whether this is what you want, but the clade models in codeml allow for more than 2 branch types. you can search for "clade" in pamlDOC.pdf and find the relevant sections.
ziheng
Nina Ramos
Mar 23, 2022, 2:37:06 PM3/23/22
to PAML discussion group
Hi all,
I have come conducted this analysis comparing (M0 and M2) to identify differences between the foreground and background. However, I had specified more than one species as the foreground, unaware that M2 is restricted to identifying just one foreground branch against the rest of the tree. I may have misinterpreted the Table 4 in Jeffares et al. (2015) that states that the hypothesis being tested in the two-ratio vs one-ratio comparison is "Are the foreground branches that you specify more likely to have different dN/dS from the background branches?"
Codeml allowed me to run this analysis to completion despite the input of multiple branches as the foreground unlike the problem Alex was met with. Is this still a valid approach or result within the output? Any input on this issue is greatly appreciated. Thank you!
I have come conducted this analysis comparing (M0 and M2) to identify differences between the foreground and background. However, I had specified more than one species as the foreground, unaware that M2 is restricted to identifying just one foreground branch against the rest of the tree. I may have misinterpreted the Table 4 in Jeffares et al. (2015) that states that the hypothesis being tested in the two-ratio vs one-ratio comparison is "Are the foreground branches that you specify more likely to have different dN/dS from the background branches?"
Codeml allowed me to run this analysis to completion despite the input of multiple branches as the foreground unlike the problem Alex was met with. Is this still a valid approach or result within the output? Any input on this issue is greatly appreciated. Thank you!
Ziheng
Apr 3, 2022, 11:34:27 AM4/3/22
to PAML discussion group
under the branch-site model, you can have only two types of branches: background branches (labeled #0) and foreground branches (labeled #1). having no label is the same as using label #0.
you can label multiple branches on the tree as the foreground branches. in other words, multiple branches can be labeled #1.
best, ziheng
Reply all
Reply to author
Forward
0 new messages