https://www.youtube.com/watch?v=mBcY3W5WgNU
*Thread Reply:* LMFAO I just realised the description is:
"Sing along to ’Spam Song' with this official karaoke style Monty Python lyric video. "
*Thread Reply:* and program is here: https://spaam-community.github.io/#/events/spaam3/programme
*Thread Reply:* I can forward to you 🙂
*Thread Reply:* For quick reference here is the link and meeting ID and passcode for the Zoom meeting: https://us02web.zoom.us/j/82064835502?pwd=Q0NzVWM3cGtnRm9Vb2pyK3Vha00zdz09 Meeting ID: 820 6483 5502 Passcode: 831478
*Thread Reply:* Most emails sent by Kelly: blevinske1@gmail.com
*Thread Reply:* (and forwarded to your york account : ))
*Thread Reply:* Thanks Kelly and James! Sorry about my disorganisation 🙄
*Thread Reply:* #SPAAM3 you mean? or the bone trafficer?
*Thread Reply:* https://twitter.com/search?q=%23SPAAM3&src=typed_query&f=live
*Thread Reply:* ah, just saw you replied with the actual hashtag 😆
https://www.nature.com/articles/s41597-021-00892-0
*Thread Reply:* A lot of grants require open access at the end of project - how will datasets then be utilised? And then from a commercial point are companies (ie ancestry companies) using this open data in some framework for profit?
*Thread Reply:* How is ‘indigenous/ethically sensitive contexts’ defined? Should e.g. Siberian sites be considered as such?
*Thread Reply:* Depends on your opinion 😉 but from my knowledge, yes, could be
*Thread Reply:* That is the main question Betsy had, not necessarily as clearly defined in Europe/Eurasia
*Thread Reply:* Yeah, I don't know how we could define it in Europe/Eurasia because there is not much discussion around it. Also the concept of "indigenous" itself inside Europe is not easy to define because there has been so much migration through Europe over the time. Just a bunch of thoughts here on the subject. But I think European scientists (including me) should try to include more local communities and people from the culture in the research process, indigenous or not indigenous.
*Thread Reply:* we use fastANI for average nucl. identity calc. it is kmer based and impressively fast. https://www.nature.com/articles/s41467-018-07641-9
Paper Nikolay shared: https://pubmed.ncbi.nlm.nih.gov/28158639/
*Thread Reply:* Here is another great way to reduce redundancy in a database https://microbiomejournal.biomedcentral.com/articles/10.1186/s40168-018-0399-2
*Thread Reply:* @Felix Key I haven't seen a formal comparison but KrakenUniq allows basically an unlimited database size while MALT does not so I would expect KrakenUniq would be at least more sensitive, not sure about specificity though
*Thread Reply:* @Felix Key I did not compare MALT vs. KrakenUniq on synthetic data but ran MALT and KrakenUniq (using database built on same reference sequences) on a few benchmark samples where we are pretty sure we know the microbes. The results delivered by both tools were mostly in agreement, so no major discrepancy
*Thread Reply:* hi marcel we have not decided on that. guess a snakemake would be an option (and easy to rerun by other ppl) but not sure if that is what is going to happen at the end.
*Thread Reply:* Thanks, but do you think it would also make sense to publish the output (e.g. a file with all IDs of included refseqs)? Otherwise I would be afraid that screening pipelines of different labs are turning into ‘black boxes’ hampering comparability and also troubleshooting.
*Thread Reply:* sure databases grow constantly and interfere with explicit reproducibility. the approach ian presented is mainly a reproducible (non-random) method how genomes are selected and it should lead to a reasonable output whatever version of, lets say refseq, is used. of course for mirroring an exact database someone can always share the IDs used.
@channel reminder: please add your labs to: https://tinyurl.com/spaam3-labs
And also 🤫 <#C02D992D5TN|spaam-pets>
*Thread Reply:* every virtual conference makes me incredibly envious of all the amazing pets:meow_party:!! All i have our semi alive plants 🪴
*Thread Reply:* plants are great too!! I love a good jungle!!
SPAAM3_Islandvibe https://gather.town/invite?token=ZfpvIs5r SPAAMiscoolio2021
SPAAM3_Geekingout https://gather.town/invite?token=lyJnL4y0 SPAAMiscool2021
*Thread Reply:* Yes please! (osteologist here 🙋♂️)
*Thread Reply:* What would be most useful to you in such a workshop?
*Thread Reply:* Oof, good question... lemme think about it
*Thread Reply:* I can also ask my lab what kind of things they would like to know
*Thread Reply:* @Abby Gancz I think the most effective sampling occurs in the field, right? So perhaps discussing the necessary equipment to wear while sampling (in an optimal scenario, but also considering practicality), as well as storage of the sample until sending it to the lab. Perhaps also a basic presentation of the potential contaminants and the effect they have on interpretation of the results.
https://www.protocols.io/view/dental-calculus-field-sampling-protocol-sabin-vers-bqecmtaw
*Thread Reply:* Yes... sort of... you can copy protocols directly to an OSF project
*Thread Reply:* I haven't explored it fully yet
*Thread Reply:* I know it's backed up on GitHub and clockss archived
*Thread Reply:* But this sounds awesome
*Thread Reply:* Where are the**
*Thread Reply:* Pretty well hidden and nothing in the docs (i think)
*Thread Reply:* OSF documented it: https://help.osf.io/hc/en-us/articles/360063021413-protocols-io
(https://bmcgenomics.biomedcentral.com/articles/10.1186/s12864-019-6355-0 - written by Troll et al. No less)
*Thread Reply:* Yes, this is the method developed by our lab at UCSC. It works well for us
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8141684/
*Thread Reply:* @Nasreen Broomand is that what you referring to?
*Thread Reply:* we started using it but still a lot of room for improvement. I got many dimers
*Thread Reply:* @Maria Lopopolo try adjusting your iPCR cycles
*Thread Reply:* We're planning to start testing this method
*Thread Reply:* @Nasreen Broomand we do a qPCR before and we are adjusting it does help a bit but not so much. I am starting to think many times depends a lot from the quality of the DNA and also I notice smaller peaks like before dimers
*Thread Reply:* what is the beads to DNA ration you use to purify after amplification? we are using beads and not the elute kit
*Thread Reply:* @Maria Lopopolo there's another way to estimate cycle number without qPCR that I've found is easier and I get better results. I'll post that and the name of the cleanup enzyme I was talking about in a bit after the talks. We use 1.5x spri.
*Thread Reply:* What QC methods (e.g. qubit, tape, etc.) do you use to quantify and at which stage of your prep?
*Thread Reply:* @Nasreen Broomand thank you very much this is precious 🙂
*Thread Reply:* Hi Maria! We do a qubit reading in the ancient lab before library prep (in a special fume hood in a separate part of the lab since the qubit standard #2 obviously has a lot of DNA and we definitely don't want to contaminate anything!) Post-iPCR cleanup we just do qubit and tapestation 🙂
*Thread Reply:* our tapestation machine broke at one point so we were using a fragment analyzer for a while instead, which also works well. I wouldn't use nanodrop in lieu of qubit though if you're relying on those numbers for anything other than pooling
*Thread Reply:* @Nikolay Oskolkov Yes we have been able to generate nice damage patterns for bovine reads. For the elephant, in Morocco it was an hypothesis that Loxodonta could have been present in Morocco and we detect it in 2 samples. So a bit surprising but not that much to retrieve some
*Thread Reply:* It happens anyway
*Thread Reply:* The problem is that there are just so many adapters in the carp genome
*Thread Reply:* It picks up anything even if you still have like 5bp left of the adapter in there
*Thread Reply:* E.g. We adapter clip with even just 1bp overlap, and still get hits
*Thread Reply:* in my case cyprinos carpio signal always disappeared after I was more careful with adapter removal
*Thread Reply:* (sorry, pizza hands)
*Thread Reply:* Huh ok. What did you do instead?
*Thread Reply:* Or extra, rather
*Thread Reply:* It also picks up adapter artifacts whereby a stubby adapter ligates to an already-ligated stubby adapter
*Thread Reply:* two (or more) rounds of adapter trimming solves that
*Thread Reply:* Hmm interesting... Maybe something to implement in eager...
*Thread Reply:* @Pete Heintzman I do agree, we also run minimum 2 rounds of adapter trimming (and sometimes from different tools)
*Thread Reply:* would running adapter removal twice, remove bases that are actually not adapters? would that mess up with removal of duplicates?
*Thread Reply:* Had carp in a sample from Troy back in 2015 (as a new aDNAer). Told myself a whole story about how the person might have been buried with fish for some religious reason and was about to go do some research...mentioned it to a senior lab member and realized what was going on 🤣.
*Thread Reply:* I actually was curious on how different our results would have been with krakenUniq — I started downloading the database last night but unfortunately didn’t get the analysis done before today
*Thread Reply:* i suspected as much 🥲: giving a talk knowing that someone will ask you "why didn't you do this thing that I too just learned that we should be doing"
*Thread Reply:* I always try to use the full NCBI NT, basically all organisms ever sequenced by human being 🙂
*Thread Reply:* something clearly had gone horribly wrong.
*Thread Reply:* I'm sure it's totally reasonable! They have been travellers, haven't they 🤪
*Thread Reply:* true @Katerina Guschanski perhaps I was sitting on the next Nature paper!! the bliss of being super green in data analysis 💚
*Thread Reply:* I once found a mammoth with malaria...
*Thread Reply:* Turned out it was PhiX contamination in a malarial genome.
*Thread Reply:* "mammoth with malaria" would make an impressive cover @Pete Heintzman - think of the media headlines :meow_party:
*Thread Reply:* I think my favorite was dolphin and cannabis in a mountain gorilla gut sample 😄
*Thread Reply:* Amazing!!! this is such a great thread for examples in teaching!
*Thread Reply:* Actually this might be a good use of KrakenUniqs hierarchical database runs
*Thread Reply:* We can discuss this James! We are going to publish this line of thinking soon (with Anders Göthreström and Love Dalen). Nf-core eager is definitely a good place to host this 🙂
*Thread Reply:* https://www.sciencedirect.com/science/article/pii/S1040618220307746?dgcid=rsssdall
*Thread Reply:* Yes! of course there is this one!
*Thread Reply:* but I got the feeling that some of the things mentioned have not been fully covered in that article.
*Thread Reply:* (Allie's figures were from that, but Indeed Nikolay had more in-depth examples Indeed)
*Thread Reply:* Also in this review: https://www.annualreviews.org/doi/abs/10.1146/annurev-micro-090817-062436 we also touch a bit, but it will be great to have a more specific publication
*Thread Reply:* It was while hearing to Nikolay that I thought about it
*Thread Reply:* Overall, I think it would make sense if there is enough material that hasn’t been mentioned in previous publications
*Thread Reply:* Thoughts @Nikolay Oskolkov and @Allie Mann?
*Thread Reply:* I would be happy to contribute!
*Thread Reply:* Maybe we can discuss this after SPAAM in one of the channels. First thing would be to identify points of interest and evaluate if all together they could make a story
*Thread Reply:* Drafting any manuscript takes lot of time, it must make sense
*Thread Reply:* We apply depth and breadth of coverage thresholds reported by KrakenUniq for eliminating obvious false-positive organisms and use all organisms reliably detected in at least one sample for building a project-specific MALT database. Ideally those thresholds should be sample-specific, we are working on it, there are some ideas but so far hard thresholds
*Thread Reply:* Do you have your workflow published or deposited somewhere for reference? We have done a similar approach but with 2 malt runs, one with a tiny database with our species of interest to filter out anything that is definitely not what we are looking for. Then extract all the aligned reads, and run them through a big Malt databse with all bacteria... but as you said, the memory usage is a big problem. Your way of flipping the approach around is much more HPC friendly
*Thread Reply:* It is going to be published soon. I think during this conference I heard a similar line of thinking a few times already, i.e. using some way of pre-selecting reliable candidates and building MALT database on them. Glad to hear that other people also thought about it
*Thread Reply:* Soooooorry for that! I’ll control myself next time… 😉
*Thread Reply:* Your child is very cute 🙂
SPAAM3_Islandvibe https://gather.town/invite?token=VrLwFkWq SPAAMiscoolio2021
Hi all, I've attached the ethics form I discussed yesterday! Feel free to reach out with any feedback/suggestions 🙂.
Day two drawings from @Petra Korlevic 😍
https://twitter.com/petrathepostdoc/status/1433507735673376771?s=19
}
Twitter
*Thread Reply:* This is absolutely awesome! How do you do it? I mean is it drawn on a computer or by hand and then digitized? You are amazing!
*Thread Reply:* There were people sprites with silly hats!
@Betsy Nelson has proof ;)
*Thread Reply:* close enough then! too bad you cant make your avatars in little bacteria
*Thread Reply:* That's an awesome idea... Should Pitch it to them!
*Thread Reply:* no but i was definitely thinking about something of the sort at some point 😄
closest thing is we have some school engagement on flying insects so a good excuse to teach kids how to draw silly bugs
*Thread Reply:* I’d sign up for a silly bug class too 🐝
*Thread Reply:* You could probably set up a patreon 🤔
*Thread Reply:* hah wonder if there would be a crowd for a sci comm style patreon 😄
*Thread Reply:* I really think there would be! There is bigger bigger outreach requirements with grants etc, and so there is interesting interest in ways of doing so.
Certainly in our department I'm in there were lots of volunteers to help out drawing stuff in the archaeological science colouring book: http://christinawarinner.com/outreach/children/adventures-in-archaeological-science/
*Thread Reply:* oh i love the colouring books 😄 though i never finished my copy of it
*Thread Reply:* Isn’t Braken preforming some type of normalization by genome size?
*Thread Reply:* If I remember well it does not, it tries to assign the reads in higher taxonomic ranks to an actual species but no genome lengths accounted
*Thread Reply:* This is really interesting! How do you account for reads that cannot be classified to the species level (and presumably then you don’t know the genome length)? Do you only focus on normalizing by those taxa that you do have good classification for?
*Thread Reply:* Yes what I do is to run Bracken and then get the reads assigned at the species level for the normalization. Reads that are not placed by Bracken (after Kraken classification) in a species remain unaccounted (unfortunately).
*Thread Reply:* We have been doing the same as @Claudio Ottoni mentioned: normalizing by genome size after Kraken. For reads that could only be assigned to a higher taxonomic level, we used median genome size for this taxonomic group (which is not ideal but in our mind still better than doing nothing). However, since starting to use Braken we stopped normalizing. Maybe @Adrian Forsythe can comment on the internal normalisation process of Braken
*Thread Reply:* Sharing input on Bracken from a former lab member (Jaelle Brealey): Bracken takes into account the kmer size used to build the kraken database and the read length, and does some fancy statistics to “derive probabilities that describe how much sequence from each genome is identical to other genomes in the database, and combine this information with the assignments for a particular sample to estimate abundance at the species level, the genus level, or above.” If they claim they estimate species abundance, not just read counts, then they should by default take into account read length, genome size, etc.
*Thread Reply:* I could perhaps comment on the first question, to my experience it is 1TB of RAM for Bowtie2 which is much less than what is needed for MALT (3.5 TB of RAM). Let us see if Maxime can confirm it 🙂
P.S. after double-checking it was 1.3TB disk space for storing Bowtie2 index, and 3.3 TB of disk space for storing MALT database
@Maxime Borry have you come across this pipeline? https://github.com/frederikseersholm/getLCA also lca from a sam file
*Thread Reply:* No problem, think the publication date was around 2016. The documentation is terrible. It uses a python script and spits out a txt file of matches and read counts. As a user and not really an understander of these tools it was quite a confusing process for me 😂 I also never managed to get it to report anything that wasn’t to sp level, but again that could very much be user error!
https://maximeborry.com/post/rocksdb/
*Thread Reply:* I thought it was about sam2lca, but maybe I misunderstood
*Thread Reply:* This is part of the "guts" os sam2lca
*Thread Reply:* Thank you so much! Thanks to the spaam community for being awesome! Love y'all!:facewithcowboy_hat:
*Thread Reply:* I awas about to say 😆
*Thread Reply:* Ha-ha, I do QC both before and after adapter removal, it is nice to see how QC metrics change after adapters have been removed
*Thread Reply:* We also do that in eager too... peoople weren't checking their QC reports with EAGER1, and when they started looking at MultiQC they were shocked at the before- and after- deifference
*Thread Reply:* In fact we do that too. But since the QC also included some metagenomic profilers and other tools, these tools are only run after AdapterRemoval, it is more accurate mentioning the QC after the AdapterRemoval step… 😉
*Thread Reply:* Nikolay! It was so nice to “meet” you. Thank you for all the incredibly helpful tips and support!
Here is my script for MSAs: https://github.com/campanam/vcf2aln
*Thread Reply:* I think a lot depends on how the pandemic and the response to it develops
Gathertown link: https://gather.town/invite?token=VrLwFkWq Password: SPAAMiscoolio2021
*Thread Reply:* A desktop computer rolled in James’ hair?
https://en.wikipedia.org/wiki/Hockey
and final day done 🙂 https://twitter.com/petrathepostdoc/status/1433861839222448152/photo/1
Twitter
*Thread Reply:* The tooth on Irina's picture looks like it has a face in it to me, and it looks like it's in bliss because it's been hugged by the calculus
*Thread Reply:* Which is perfect because of @irinavelsko's Twitter handle
*Thread Reply:* https://twitter.com/FzzyToothSweatr?s=09
*Thread Reply:* ahaha i drew one of these cozy teeth for a talk on isba and just fell in love with how silly and friendly it looked
*Thread Reply:* They are amazing 😍
*Thread Reply:* I also thought all the teeth are smiling, they’re so cute!
(even at small resolution the file is quite giant)
*Thread Reply:* That's beautiful 😍
*Thread Reply:* Thank you so much for all your effort into this, it really makes these remote events to have much more impact 🤩
*Thread Reply:* glad it added a twist on the zoom meeting 😄
*Thread Reply:* Thank youuuu!!! This is wonderful!
*Thread Reply:* Would you mind @Petra Korlevic if we print this as a poster??!
*Thread Reply:* absolutely fine, i sent the link to the A0 file to to @Miriam Bravo but here is the google drive link for anyone that wants a version (file has 21MB... dont know if slack can handle that) https://drive.google.com/file/d/1Zv_i17gHQBDhwq2Weq5Q15Mheufx5_b2/view?usp=sharing
*Thread Reply:* Thank you so much @Petra Korlevic!
*Thread Reply:* I am speechless, it’s amazing @Petra Korlevic!
@Petra Korlevic’s amazing poster in physical form! (Apologies for my terrible photography)
*Thread Reply:* it sure adds a lot of color to the area 😄
*Thread Reply:* It really does!
*Thread Reply:* Need to get more of your doodles and stick them up everywhere 😉
*Thread Reply:* well there used to be some over in the "methods" corner since i was sitting there and doodling constantly (not sure what it is now or who sits there but it's around Janet's area)
*Thread Reply:* oooh I might go and do a hunt later and see if I can find them. They did take out a lot of stuff out htough
Hi @channel ! We designed some SPAAM stickers and we will send one to you if you are interested and participated in SPAAM3. Please get in contact with me, and we can arrange the sending by post of the sticker!
