Hello
In phylogenetic analysis:
What is the purpose of misidentified or contaminated sequences in a data set?
How can I remove this contamination?
If I use a Neighbor-Joining algorithm (with a K2P model) in MEGA, how can I recognize the misidentified sequences?
Hi,
misidentified or contanimated sequences should not be included in a dataset: they have NO purpose but represent some sort of error.
How you identify these is a difficult question. If you geenrated the data yourself, you need to ask yourself: Did I identify all my species correctly? If not these may be such misidentified species. If you rely on published data, misidentified species / sequences is very difficult. They may appear at unexpected positions in your phylogenetic tree or may feature unexpected large numbers of mutations or may be identitical to sequences of other species (but of course there may also be good biological explanations to any such features!).
Hi,
misidentified or contanimated sequences should not be included in a dataset: they have NO purpose but represent some sort of error.
How you identify these is a difficult question. If you geenrated the data yourself, you need to ask yourself: Did I identify all my species correctly? If not these may be such misidentified species. If you rely on published data, misidentified species / sequences is very difficult. They may appear at unexpected positions in your phylogenetic tree or may feature unexpected large numbers of mutations or may be identitical to sequences of other species (but of course there may also be good biological explanations to any such features!).
As Martin said, misidentified sequences have no purpose, they are an error. On the question about how to identify them, I would rely on tree anomalies, i.e. sequences appearing in unexpected positions on a tree (either taxonomically or without much biological sense - clades well known to be different for that gene that cluster together for example...); problems of long branch attraction and low support values; and so on.
Just look at your aligned sequences. If any of them is too much different from the others, then you spotted it. Indeed, "too much different" might turn out tricky in some instances.
If you still have doubts, BLAST your sequences. Sequencing from cloned fragments sometimes leads to bacterial (plasmid) reads.
Have a nice day
- Badges
- Science method
- Similar topics
- Biological Science
- Evolutionary Biology
- Phylogenetics
- Phylogenetic Analysis