Becky needs a review on: https://github.com/SPAAM-workshop/AncientMetagenomeDir/pull/162
However I will need to add to our spaam github organisation. If you send me your username I will add it, and they Becky can take it from there!
}
Thanks Pete! Great to hear from you again.
Actually, I've just added Smith 2015, which was much simpler than the Slon samples and would make an easier first review 🙂 https://github.com/SPAAM-workshop/AncientMetagenomeDir/pull/167
> You've invited Pete Heintzman to SPAAM-workshop! They'll be receiving an email shortly. They can also visit https://github.com/SPAAM-workshop to accept the invitation.
First Issue to merge I didn't have to do anything 😍 thanks guys
*Thread Reply:* Yeah - that was my bad. For context, the final ages were calculated late on in the work, but the sample depths (which we using to cross-correlate all proxies) remained static throughout. re. FigShare: and vice versa, so probably ok to just cite the publication DOI if only one DOI is desirable.
*Thread Reply:* I found this with the Armbrecht paper: depth was given for all the samples but age was only given for the deepest sample. Since all the samples were recent I was able to put down 100 years for the shallower samples too, but not sure what I would have done otherwise.
*Thread Reply:* Threw me for a loop but very grateful for @Pete Heintzman helping me link IDs to ages!!
*Thread Reply:* I think this may be the way to go.
*Thread Reply:* I don’t know if this is possible, but can we create this field with the github actions?
*Thread Reply:* What do you mean? I mean the person submitting must work that out
*Thread Reply:* You have to take it from the publication. A lot of people don't give sample alias/library names in the ENA/SRA submission
*Thread Reply:* I mean if we have the columns defined, when doing the PR if we can have a small function that combines the different fields
*Thread Reply:* @channel: here is a good primer on age-depth modelling used for dating sedimentary sequences:
*Thread Reply:* Btw, you could always make a draft PR, so you once everyone agrees we can sent it straight in
*Thread Reply:* And some tweaks (which is why I suggest the draft PR ;)):
```## sedimentary_sequence
- Sediment only
- Identifier for sequence sample was taken from, e.g. "core3", or zonea19
- Typically cores, or quadrant of excavation
If not reported, NA```
*Thread Reply:* ```## sample_name
- Unique identifier for that sample as used in the publication
- If samples are referred to by multiple names, use the most informative
- If samples cannot be *directly* linked to data files by any names in the publication, generate names in the format e.g. [sequence][depth][original name]
> ⚠️ Mandatory value```
*Thread Reply:* otherwise I'm happy with it!
*Thread Reply:* Agree - this looks good!
*Thread Reply:* @Becky Cribdon time for a PR! We will need to work out how to retroactively go back through published papers though...
*Thread Reply:* Changes noted and draft PR created 🙂
*Thread Reply:* Sure… but i have no idea how to that. Pointers?
*Thread Reply:* you don't need to do anything
*Thread Reply:* @Becky Cribdon @Pete Heintzman made you a new team. So in the future people on submitting environmental PRs can go @thedir-team-dirt for reviewers or help, and you both get notificaitons.
*Thread Reply:* Heh heh, team dirt. I like it.
*Thread Reply:* Shame it's only us three.
*Thread Reply:* WE will get thre 💪
*Thread Reply:* I'm wondering if I should submit a second abstract to ISBA about this...
*Thread Reply:* And if there are any sedaDNA conferences feel free to present there about this!
*Thread Reply:* Abstract about what precisely?
*Thread Reply:* Tell people it's here, what it does, etc
*Thread Reply:* Want to recruit more contributors
Thoughts https://github.com/SPAAM-workshop/AncientMetagenomeDir/pull/268#issuecomment-688640430?
We can add this to Becky's restructuring PR if you guys agree
}
ah and @Anneke ter Schure just approved the Braadbaart paper just as Iwrote that message, awesome! You can merge now @Becky Cribdon! https://github.com/SPAAM-workshop/AncientMetagenomeDir/pull/268
Done, and issue made: https://github.com/SPAAM-workshop/AncientMetagenomeDir/issues/284
}
*Thread Reply:* Maybe it's relevant if there's been some selection or enrichment before generating the data?
*Thread Reply:* But I'm imagining a typical user wanting to find comparable datasets, not really the analysis methods (i.e. what alignments).
*Thread Reply:* But, if the data itself is taxonomically limited, maybe it would be helpful to know that.
*Thread Reply:* Yeah I was also thinking that
*Thread Reply:* like mt data from Slon et al. (not that we specify the data, but they may only upload data along those lines 😏 )
*Thread Reply:* So we have a tentative yes from yuo?
*Thread Reply:* I agree with Becky's suggestion, although it is slightly different from what the reviewer is specifically asking.
*Thread Reply:* tl;dr: I don't think the reviewer's suggestion in necessary, otherwise we would need to update every time someone reanalyzes the data.
*Thread Reply:* In my specification I have stated 'the original purpose' of the data.
*Thread Reply:* Bceause I agree that would come complicated, but assuming it's the same DNA extraction that I think is the main influence you would have downstream
*Thread Reply:* I guess it makes sense, for example people looking for just hominin DNA might do some funky stuff in the future to try and enrich for that
*Thread Reply:* Or if someone did some fine-filtering of the sediment to remove certain things (that might bias results)
*Thread Reply:* There are some seriously deep rabbit holes if one starts going into differences in wet lab methods used too deeply. As a compromise, I think enrichment v. not of libraries is the most important.
*Thread Reply:* Yeah sorry, I actually meant that by just saying 'they looked at animals or bacteria' that you can ignore that whole thing of lab-methods completely.
By using eukaryotic vs microbial, I would hope and assume people who are interested in animal DNA will do similar things (and would only download that data), and people who are only interested in microbial DNA would do similar things and look for that data...
*Thread Reply:* Perhaps use NCBI taxonIDs as these will offer the most flexibility (from single species to phyla)?
*Thread Reply:* I looked at that actually! And while it was a good idea at first, I observed just now that people filtered in lots of weird an wonderful ways... which would result in very long lists. So far all the papers could be split into: faunal, floral, or microbial (or some combination of the three), in my opinio nat least.
*Thread Reply:* A core drive is a section of a sediment core. Some coring methods recover one drive (e.g. 1 m) of a core at a time from the same hole. Multiple holes are cored so that there are overlapping drives to reconstruct a composite core record (as the ends of drives tend to be disturbed).
The details in your example: 1B-1B_23-25 1 = coring site on the lake (in this case, centre) B = hole code 1 = core drive/section B = core section is cut in half longitudinally. Half B was sampled for DNA 23-25 = depth in cm into the drive that DNA sample was derived from
*Thread Reply:* Ahh cool thank you
*Thread Reply:* Personally, I would rather not specify "study type" if it hasn't affected the raw data the 'Dir is pointing to.
How about changing the name from "study type" to something like "enrichment", to mark whether the sample has been manipulated to change the taxonomic profile? So, if they enriched for mammal DNA it would be "faunal"?
But, for example, Gaffney 2020 was a floral study, but the raw data contains everything. We just picked out the plants for further analysis. So I wouldn't want to specify a "study type" for that: that could suggest the data is more limited than it is.
*Thread Reply:* Yeah good point
*Thread Reply:* I've not looked specifically if any filtering was done though
*Thread Reply:* So we need a new title 🤔 .
*Thread Reply:* taxonomic_category?
*Thread Reply:* originaltaxonomicpurpose?
*Thread Reply:* Could you expand on "I've not looked specifically if any filtering was done"? What do you mean by filtering here?
*Thread Reply:* As far as I understood the reviewer:
" A general thought on the environmental metagenomics category (most closely aligned with my expertise): a column indicating the taxonomic focus of the published study might be useful, so that there is some indication of what alignments the authors performed. "
he wants to find comparative datasets e.g. if he is interested in looking at e.g. floral eDNA from different sites at the same time period
*Thread Reply:* (I say he as I'm pretty sure it's Mike Bunce who reviewed 😉 )
*Thread Reply:* But he wants to comapre results of the original paper
*Thread Reply:* I am assuming that this implies that he would assume there were upstream 'decisions' that may influence extraction strategy for example.
*Thread Reply:* So he wants to make sure these are approximately like-by-like.
what I meant in my message above, is that the data I have input into the PR under the current column study_type, I've just done it based on what the originally publication was focusing on
*Thread Reply:* I have no actually looked if there was any enrichment
*Thread Reply:* (sorry if this still isn't coming off as clear 🤦
*Thread Reply:* What about just study_taxonomic_focus?
*Thread Reply:* Sorry, I'm finding this really complicated! But thanks for clarifying your statement.
So, for the reviewer, he would be searching the 'Dir for floral data and wants a column that would let him exclude a study that enriched for mammal DNA, because there's unlikely to be any floral DNA for him to use? I agree that would be useful.
But I think "taxonomic focus" or similar are still too broad: having a taxonomic focus doesn't necessarily mean the data is taxonomically limited. His assumption that upstream influences downstream wouldn't always hold, and could be misleading.
If I wanted mammal DNA and Gaffney 2020 had a "focus" of plants, would that suggest I exclude it? Despite it actually having raw data for everything?
Is there a title that implies actual limitation only, so studies like Gaffney 2020 could be "none" or "NA"?
*Thread Reply:* No don't worry. I'm also finding it complicated. To be clear: I am assuming that he may wonder if there is enrichment has happened, but I don't know for sure. I think I may have confused both you and @Pete Heintzman about mentioning enrichment 🤦 . We shouldn't focus on that (I was reading about TARA oceans the other day and they did particle filtering so maybe that's why I brought it up).
Looking through the comments, what he might actually be thinking about is 'meta-analysis' (which he says is something important that we are missing from the 'benefits' of the project). So he's sort of looking for some basically annotated bibligraphic info, maybe?
In which case it's actually a bit detached from the purpose of the project but I understand him from that point of view.
*Thread Reply:* So to take your example above:
> So, for the reviewer, he would be searching the 'Dir for floral data and wants a column that would let him exclude a study that enriched for mammal DNA, because there's unlikely to be any floral DNA for him to use? I agree that would be useful. Rather than enrichment, maybe what he means is that he's look for other studies that looked at flora. So he can then take their already publisehd 'OTU tables' (so to say), to compare with his own?
*Thread Reply:* Riiight, so instead of a flag for limitations, this column would be like a flag for particular detail?
*Thread Reply:* yeah sort of. What was the original purpose for the generation of that particular data
*Thread Reply:* @James Fellows Yates’s last comment is how I interpret the reviewer’s comment. However, having a separate column for shotgun/enrichment would also be very valuable to inform re-analysis of the data sets. (starting a new topic to continue this)
*Thread Reply:* Yes, I prefer "primary". Thanks for clarifying this above.
*Thread Reply:* Ok, will update
*Thread Reply:* By the way, the sequence data for the new samples, that are not in Ahmed2018, do not seem to have been made available.
*Thread Reply:* Looks legit 🙂 Approved and merged.
Y/N? https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7643826/
*Thread Reply:* Y. For the bottom sediment samples only.
@Pete Heintzman https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7887274/ one for 'the Dir?
Need one approval by someone here @channel!
https://github.com/SPAAM-community/AncientMetagenomeDir/pull/377
}
*Thread Reply:* Go ahead and merge both!
*Thread Reply:* Still needs a 'review' 😉
*Thread Reply:* Oh pants. One sec
*Thread Reply:* Hmm ok. We can keep an eye on it then. I guess we do have precedence with the pathogen stuff, but that stuff is often more clear cut because each paper is basically referring to a single taxon, but it might be more variable with sedaDNA...
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8208683/ one to include?
@channel y/n? https://www.ncbi.nlm.nih.gov/labs/pmc/articles/PMC8257590/
*Thread Reply:* Would say yes, although I see that all raw reads are combined into a single accession/fastq. Is there a policy on this? @James Fellows Yates
*Thread Reply:* The read IDs should allow you separate i think...?
Or is it multiple samples merged into one?
*Thread Reply:* Multiple samples merged into one FASTQ... as far as I can tell.
*Thread Reply:* That's really shitty
(if you wanted to check the metadata anyway, the PR is here: https://github.com/SPAAM-community/AncientMetagenomeDir/pull/431, you can see under 'files changed')
Yes, I read that Vernot in his paper adapted Kallisto to seda DNA and works very well. :)
}
@Barbara @James Fellows Yates Am reviewing Moguel2021, but need some clarifications: • Samples S1 and S2 look to be modern top sediments, and so should not be included in the 'dir? • Samples 501 and 502 are <6000 years in the pub. but 3000 yrs in the 'dir. Where did the 3 ka ages come from? I note that these samples are from different depths, so may not be the same age? • There is a discrepancy on ENA: SRS4426892 and SRS4426891 are either sample 501 or S2, depending on whether one looks at the experiment or sample alias. Which is which?
*Thread Reply:* I thought we said for sequence cores we kept also modern?
I took the dates ftom the mid point of one of the schematic diagrams
*Thread Reply:* For the last one of if mismatch is ENA then that's up to Barbara
*Thread Reply:* Thanks for the quick clarification, James!
*Thread Reply:* Please check I got the dates right though...
*Thread Reply:* Or you make the same rule-of-thumb interpretation
In case anyone was thinking of adding the new Wang et al. 2021 dataset — the uploaded data are the aligned reads only and not raw data. So not one for the AncientMetagenomeDir. https://www.nature.com/articles/s41586-021-04016-x
*Thread Reply:* Then we should note it down for the MIxS checklist
*Thread Reply:* Please post it in <#C01BX7EM4EL|metadata-standards> so we don't forget!
New addition for the ’dir, although the raw data do not seem to be available yet (may change in the coming days): https://www.ncbi.nlm.nih.gov/bioproject/799375 https://www.sciencedirect.com/science/article/pii/S0277379122000191
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9132974/
*Thread Reply:* I quickly had a look at it. I think what they are saying is 'the off-target' DNA from the captured libraries are similar to shotgun ones in terms of microbes.
*Thread Reply:* but need to read it fully, I might have understood wrong
*Thread Reply:* Where do they mention the off-target?
*Thread Reply:* Abstract: 'These data are consistent among study sites and between replicates processed with different methodologies (shotgun sequencing and targeted capture), which highlights that the “off-target” fraction of metagenomic data used to study macro-ecosystems can also be used to investigate synchronous changes in microbial communities.'
4.1 Structural and functional shifts in microbial communities '...three different sequencing targets—shotgun sequencing, PalaeoChip Arctic v1.0, and a Bovidae specific bait-set (Murchie, Monteath, et al. (2021); Murchie, Kuch, et al. (2021); Murchie et al. (2022))—all biological replicates cluster together (Figure 2; Figures S2 and S8).'
*Thread Reply:* Jesus fucking Christ the methods are so bad then?!?
*Thread Reply:* Well... @Kadir Toykan Özdoğan willing to make an issue on the Dir 😅
