SEA-PHAGES | Comparative analysis

Link to this post \| posted 09 Mar, 2021 20:53
susannepfeifer	Dear SEA-PHAGES team, If time allows, we would like to perform a small comparative analysis this semester. As we are planning to submit a manuscript with our students, this raises the question whether we i) can include any phage that has been deposited in GenBank (by citing its accession number) or ii) should limit ourselves to comparisons with phages that have already been published in a peer-reviewed journal? Any advise would be greatly appreciated! Best, Susanne

Link to this post | posted 09 Mar, 2021 20:53

Dear SEA-PHAGES team,

If time allows, we would like to perform a small comparative analysis this semester. As we are planning to submit a manuscript with our students, this raises the question whether we i) can include any phage that has been deposited in GenBank (by citing its accession number) or ii) should limit ourselves to comparisons with phages that have already been published in a peer-reviewed journal? Any advise would be greatly appreciated!

Best,
Susanne

Link to this post \| posted 10 Mar, 2021 18:03
cdshaffer	to me this is mostly an issue of good experimental design and picking the right dataset for your experiments. So consider these two possible sentences you could write in a hypothetical paper and decide which one would be better at convincing the reviewer to accept your paper and its conclusions: 1. "To cast a wide net and compare as many as possible we collected and analyzed all phage in genbank" 2. "To ensure the highest quality of data and gene annotations we selected only phage genomes which have been published in peer-reviewed journals" if I were a reviewer I would be fine with option 1 if the study was mostly about sequence variation with little or no input from annotations. The more your experimental conclusions rely on annotations the more I would favor option 2. As a middle ground, you could also consider all the phage in the refseq database instead of all of genbank.

Link to this post | posted 10 Mar, 2021 18:03

cdshaffer

to me this is mostly an issue of good experimental design and picking the right dataset for your experiments. So consider these two possible sentences you could write in a hypothetical paper and decide which one would be better at convincing the reviewer to accept your paper and its conclusions:

1. "To cast a wide net and compare as many as possible we collected and analyzed all phage in genbank"

2. "To ensure the highest quality of data and gene annotations we selected only phage genomes which have been published in peer-reviewed journals"

if I were a reviewer I would be fine with option 1 if the study was mostly about sequence variation with little or no input from annotations. The more your experimental conclusions rely on annotations the more I would favor option 2. As a middle ground, you could also consider all the phage in the refseq database instead of all of genbank.

Link to this post \| posted 10 Mar, 2021 18:29
susannepfeifer	That's a good point though all SEA-PHAGES phages in GenBank should have undergone the same rigorous gene annotation, right?

Link to this post \| posted 11 Mar, 2021 17:39
cdshaffer	Yes I agree, I was not trying to imply otherwise. My thought experiment was more with the idea of a outside reviewer. There are many papers out there that talk about how poor the annotations are in genbank as a whole, so I could imagine a naive reviewer liking the "published set" over and above the "all genbank set". This brings up the point that one might pick all SEA-PHAGES as a test set over all phages in genbank or all published phage if annotation consistency was important to the analysis. Again, the experimenter should pick the best possible dataset for the question.

Link to this post | posted 11 Mar, 2021 17:39

cdshaffer

Yes I agree, I was not trying to imply otherwise. My thought experiment was more with the idea of a outside reviewer. There are many papers out there that talk about how poor the annotations are in genbank as a whole, so I could imagine a naive reviewer liking the "published set" over and above the "all genbank set".

This brings up the point that one might pick all SEA-PHAGES as a test set over all phages in genbank or all published phage if annotation consistency was important to the analysis. Again, the experimenter should pick the best possible dataset for the question.

Link to this post \| posted 17 Mar, 2021 17:26
susannepfeifer	Thank you - that's really helpful!

Recent Activity

Comparative analysis