Secondary structures have high conservation compared to DNA sequences. It will be difficult to assign lower ranking taxa ex genus but class or phylum may work but they are not useful as such
Your video is too good for a newbie in the field, which I am. So thanks a tonne. by the way, have you published any paper on the nearest neighbor approach which you briefly described? 'coz I didn't understand it fully :(
Hello Dan Knights! I just want to ask, which method of taxonomy assignment is the most accurate, and also which database is also the most accurate.. Accuracy has been my concern because recently I was working with my V3-V4 16S rRNA gene sequences. I was using QIIME greengenes alignment method to assign taxonomy, so my Otu2 was assigned to the family Pseudoalteromonaceae, however, when I pass the Otu2 sequence to nr/nt database in NCBI Blast, it is assigned to the Vibrio parahaemolyticus, Vibrio campbelli (anyways, the list inclines towards the Vibrio genus). This is a problem since the 2 taxa diverge even at the the order level (Pseudoalteromonaceae -> Alteromonadales; Vibrio -> Vibrionales). I compared between the sequences of Pseudoalteromonas carrageenovora (representing Pseudoalteromonas), Vibrio campbelli (Vibrio) and Otu 2: Otu2 vs V. campbelli = 100.0% Otu2 vs P. carrageenovora = 89.9% V. campbelli vs P. caraageenovora = 89.9% We can be confident that using the V3-V4 sequence, Otu2 is probably a Vibrio, which means the greengenes database has some problems. With my V3-V4 sequences, can you advise me on what way can I improve my taxonomy assignment, because I havent yet gone through my other Otu sequences, who knows the taxonomy assignment can all be messed up. Thank you.. By the way, your video is a life saver, thank you sensei :)
May I ask you a question? What do you mean by "a random subset of 1/8 of 8-mers"? Say bootstrap iteration number B=100, does that mean we will have 100 "1-mer" (8*1/8=1)?
Actually the teacher wants to say that For each taxonomic assignment a particular confidence level is given. Choose random subset of k-mers in a particular genus and it repeats the classification. Then you just count up how often the same with those random subset. The confidence is report as fraction like 0.8 or 1/8. so typically run the RDP at 80 percent confidence level or above 80 percent. The confidence level is already set so you do not need to be worry about that.
Just a small correction: the publication cited at 0:51 is Yarza et al. Nat Rev Microbiol 2014, not 2012.
YOU ARE GREAT@DAN
Secondary structures have high conservation compared to DNA sequences. It will be difficult to assign lower ranking taxa ex genus but class or phylum may work but they are not useful as such
Your video is too good for a newbie in the field, which I am. So thanks a tonne. by the way, have you published any paper on the nearest neighbor approach which you briefly described? 'coz I didn't understand it fully :(
Thanks!!
Does anyone know the reference of the figure that we can see at 2:35?
SUPPLEMENTARY INFORMATION on HMP Project. Page 27.
www.vitamindwiki.com/tiki-download_wiki_attachment.php?attId=1396
or www.nature.com/nature/journal/v486/n7402/extref/nature11209-s1.pdf
to be exact
thanks!
Hello Dan Knights!
I just want to ask, which method of taxonomy assignment is the most accurate, and also which database is also the most accurate..
Accuracy has been my concern because recently I was working with my V3-V4 16S rRNA gene sequences. I was using QIIME greengenes alignment method to assign taxonomy, so my Otu2 was assigned to the family Pseudoalteromonaceae, however, when I pass the Otu2 sequence to nr/nt database in NCBI Blast, it is assigned to the Vibrio parahaemolyticus, Vibrio campbelli (anyways, the list inclines towards the Vibrio genus). This is a problem since the 2 taxa diverge even at the the order level (Pseudoalteromonaceae -> Alteromonadales; Vibrio -> Vibrionales).
I compared between the sequences of Pseudoalteromonas carrageenovora (representing Pseudoalteromonas), Vibrio campbelli (Vibrio) and Otu 2:
Otu2 vs V. campbelli = 100.0%
Otu2 vs P. carrageenovora = 89.9%
V. campbelli vs P. caraageenovora = 89.9%
We can be confident that using the V3-V4 sequence, Otu2 is probably a Vibrio, which means the greengenes database has some problems. With my V3-V4 sequences, can you advise me on what way can I improve my taxonomy assignment, because I havent yet gone through my other Otu sequences, who knows the taxonomy assignment can all be messed up.
Thank you..
By the way, your video is a life saver, thank you sensei :)
You can use the recent version of Greengenes for 16sRNA genes
May I ask you a question? What do you mean by "a random subset of 1/8 of 8-mers"? Say bootstrap iteration number B=100, does that mean we will have 100 "1-mer" (8*1/8=1)?
Actually the teacher wants to say that For each taxonomic assignment a particular confidence level is given. Choose random subset of k-mers in a particular genus and it repeats the classification. Then you just count up how often the same with those random subset. The confidence is report as fraction like 0.8 or 1/8. so typically run the RDP at 80 percent confidence level or above 80 percent. The confidence level is already set so you do not need to be worry about that.
Help a lot on my exam!
Thank you.