Sign in

Steven Robbins

@stevenjrobbins.bsky.social
3.6K followers 551 following 1K posts

Do my science @ace_uq studying coral reef microbiomes. Data wrangler, meta-omics and long-read wonk, clean energy enthusiast, Saganist zealot, collector of weird zoology facts, other nonsense.

PostsRepliesMedia
Reposted by Steven Robbins
Christian Kost @kostchristian.bsky.social · 24/08/2026
Very happy that our paper Obligate cross-feeding of metabolites is common in soil microbial communities just came out in Nature Microbiology. See here 👇 Paywalled version: www.nature.com/articles/s41... Free read-only version: rdcu.be/fBHAb
nature.com
Obligate cross-feeding of metabolites is common in soil microbial communities - Nature Microbiology
Cultivation-dependent techniques, computational analyses and genome-scale metabolic models show widespread amino acid auxotrophies, suggesting that soil microorganisms exist within integrated ecologic...
814256
Reposted by Steven Robbins
Christian R Voolstra @reefgenomics.bsky.social · 12/08/2026
very excited about our new article: Can evolution keep up with the Anthropocene? .. or better: how can evolution keep up? we propose a predictive framework to begin to understand what is required to rebuild biodiversity! @uni-konstanz.de
085
Reposted by Steven Robbins
Ove Øyås 🇺🇦 @oveoyas.bsky.social · 11/08/2026
www.biorxiv.org/content/10.6...
biorxiv.org
The MiDAS global genome catalog: 53,501 long-read MAGs representing all core prokaryotic genera in the global activated sludge microbiome
Wastewater treatment relies on complex microbial communities, yet existing genome-resolved references for this essential engineered ecosystem remain dominated by short-read assemblies, limiting genome...
073
Reposted by Steven Robbins
Molecular Cell @cp-molcell.bsky.social · 07/08/2026
Broken chromosomes don’t stay home: Tunneling nanotubes carry damaged DNA between human cells
dlvr.it
Broken chromosomes don’t stay home: Tunneling nanotubes carry damaged DNA between human cells
In a recent Cell paper, Maurais et al. show that genomic instability drives human cells to transfer fragmented chromosomes to neighbors via tunneling nanotubes (TNTs), with heritable functional consequences, raising fundamental questions about intercellular communication, genome surveillance, and cancer evolution.
0107
Reposted by Steven Robbins
Bahtiyar Yilmaz @bahtiyil.bsky.social · 03/08/2026
A stool transplant delivers billions of microbes into a sick gut. Almost none move in for good. We followed donor-recipient pairs from whole communities down to single nucleotides, and found the gut runs a ruthless strain-level filter. Out now in Cell Reports 🧵 www.cell.com/cell-reports...
cell.com
Strain-level ecological filtering governs microbial colonization of the human gut
Baertschi et al. demonstrate that microbiota transplantation functions as a selective ecological filter. Although donor species frequently appear, only specific lineages achieve stable strain integrat...
25124
Steven Robbins @stevenjrobbins.bsky.social · 03/08/2026
More like pass a law like the cookie laws europe passed stipulating that you have to opt in for them to use your data like that. That makes more sense.
020
Reposted by Steven Robbins
Simon Roux @simrouxvirus.bsky.social · 31/07/2026
Session proposals are now open for ASM Microbe 2027 ! asm.org/events/asm-m... If there is a specific topic or speaker you would like to see in Chicago, this is your chance (until Aug 12) ! This is for Micro Symposia and Hub Sessions, but my DMs are open if you have ideas for bigger AEM sessions
asm.org
ASM Microbe | Overview
ASM Microbe showcases the best microbial sciences in the world and provides a one-of-a-kind forum to explore everything from basic microbiology to translation and application.
0510
Steven Robbins @stevenjrobbins.bsky.social · 29/07/2026
100%. If I were looking for BGCs, MAGs w/ full 16S, HQ MAGs in general, GBR-MGD would be the way to go (we hope). Plus, crux of the paper was that you can’t get many MAGs for many taxa using short reads, so they’re less likely to have HQ representatives in short read DBs, much less with intact BGCs.
010
Reposted by Steven Robbins
Nature @nature.com · 29/07/2026
Nature research paper: Diverse bacterial pattern recognition receptors sense the core phage proteome go.nature.com/45fpXCZ
go.nature.com
Diverse bacterial pattern recognition receptors sense the core phage proteome - Nature
Systematic analysis of prokaryotic STAND NTPases — relatives of animal and plant immune receptors — uncovers diverse antiviral sensors that detect most of the core structural and replicative proteins of bacteriophages.
16430
Steven Robbins @stevenjrobbins.bsky.social · 29/07/2026
which feels like an important step to more easily study marine communities, since SAGs are a pain, mostly pretty incomplete, leave out the free viruses, euks, etc. Hats off to the Aalborg team since you guys were the ones that inspired me to try ONT when we knew Illumina wouldn’t work.
110
Steven Robbins @stevenjrobbins.bsky.social · 29/07/2026
Thanks Anders! We’re certainly not the first to use long reads on marine samples, and others have circumvented these issues with short reads by using flow sorting/SAGs, but I do think we’re the first to create such a comprehensive database of marine long-read MAGs, viruses, euks…
110
Steven Robbins @stevenjrobbins.bsky.social · 24/07/2026
In theory, CheckM2 ML models should have to see complete genomes from these lineages and then should work better on them. Not sure how much of a pain it would be to do an update, though.
000
Steven Robbins @stevenjrobbins.bsky.social · 24/07/2026
Oh interesting, sorry, when I quickly searched for CheckM in the paper to see which one was used, I only saw CheckM1. I'd actually be quite curious to know what lineages CheckM2 was underperforming on for you. Feel like at some point we should shoot a list to the CheckM team so they can address it.
100
Steven Robbins @stevenjrobbins.bsky.social · 24/07/2026
Hey Thomas, congrats on your study, too! The CheckM part is interesting problem that the CheckM authors would readily acknowledge. Also the reason they released CheckM2 to address this issue, which does perform much better--big/diverse tree though, hard to get everything 100% right all the time.
100
Reposted by Steven Robbins
Australian Centre for Ecogenomics (ACE) @ace-uq.bsky.social · 23/07/2026
New work led by @stevenjrobbins.bsky.social @markoterzin.bsky.social finally published - we now have a reference of microbial genomic resources from the Great Barrier Reef. The paper came out just hours after the work presented by Marko at #ICRS2026 @icrs.bsky.social See thread here for more detail
0167
Steven Robbins @stevenjrobbins.bsky.social · 23/07/2026
Thanks, Alex! Sorry you posted your metagenomic depth-series paper recently as well, so congrats! I know this wasn't the main point, but a cool thing to see: bsky.app/profile/alex...
110
Steven Robbins @stevenjrobbins.bsky.social · 23/07/2026
Thanks Amin! Hope you're well. :-)
010
Steven Robbins @stevenjrobbins.bsky.social · 23/07/2026
Of course, this was a huge effort by all authors: @markoterzin.bsky.social, Katherine Dougan, @julianzaugg.bsky.social, Sara Bell, @patricklaffy.bsky.social, @pam-engelberts.bsky.social, Kim-Anh Le Cao, Renee Gruber, Nicole Webster, David Bourne, @xrefugee13.bsky.social, and Yun Kit Yeoh.
010
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
Sub-thread 3: Seawater microbial communities reflect the impacts of reef management strategy. bsky.app/profile/stev...
000
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
Interestingly, the microbes most predictive of these effects are the "Ghost Taxa" above.
000
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
Leveraging the GBR-MGD, we find that seawater microbes reflect and are predictive of the impacts of reef management strategies. However, there's a lot more work to be done to determine why and how this occurs. bsky.app/profile/stev...
110
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
Thanks Paul! Let me know what you think! :-)
000
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
Agreed!
000
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
Sub-thread 2: A 3rd reason for the existence of short read Ghost Taxa is that both CheckM1 and 2 still systematically underestimated completeness of some marine orders and classes. CheckM2 is a marked improvement, but still some areas of the tree to improve. bsky.app/profile/stev...
100
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
Sub-thread 1 🧵: Short-read metagenomic sequencing cannot recover genomes from many abundant marine prokaryotes due to high strain heterogeneity and platform-specific GC bias (likely viruses, too), but Nanopore long reads can address this. bsky.app/profile/stev...
100
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
Given the difference in recovery rates between short reads and our @nanoporetech.com data, and how much diversity short reads miss out on, our conclusion at this stage is that long read sequencing is a critical tool for seawater metagenomics.
020
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
We noticed that another feature common to all our Ghost Taxa was low-GC (<40%). Graphing GC and Straininess together, we find a region of the graph at the intersection of high strain diversity and low-GC where short reads seem not to venture. And we see that our Ghost Taxa sit in that region.
100
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
We've benchmarked short read and Nanopore assemblies to show that many families and whole phyla can be recovered using Nanopore long reads but not short reads. But why? We show that yes, Ghost Taxa have high strain diversity, and no, short reads mostly can't handle it...except when sometimes it can?
100
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
This applied not only to Pelagibacter, a few of which are circular/complete here, but other interesting taxa like Parasynechococcus species sp002724845, by far the most abundant prokaryote in the GBR-MGD. Notably, species sp002724845 could not be recovered via illumina-only metagenomic sequencing.
100
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
I decided that if this was the true bottleneck to MAG recovery, we should try using the longest reads possible to span strain-variable regions, make assembly trivial--same puzzle, really big and less pieces. We threw our samples on a Promethion and those taxa we worried we couldn't recover, we did.
100
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
Studies (Delmont and Meren 2019) have shown that dominant taxa like Pelagibacter have incredibly high genomic diversity--i.e. many strains in a single droplet of seawater--and logic follows that short read assemblies, having to walk many diverging paths in an assembly graph, would simply break.
100
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
been aware for many years that dominant taxa like Pleagibacter, SAR86, Prochlorococcus are few and far between in MAG databases. They're everywhere, you may get a few low-quality MAGs, but 16S would tell you that you should have buckets of them that never materialize. I'm calling these "Ghost Taxa."
100
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
The end-goal of the GBR-MGD was to identify microbes diagnostic of reef health. But this first required us to assemble a holistic DB of HQ genomes from Australia’s Great Barrier Reef--our goal was 10,000. That was a task I knew not to be trivial. Those in the marine metagenomics field have...
101
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
This is NOT a criticism of a tool, CheckM2 works shockingly well across the tree. These results are meant to highlight how useful long-read circular MAGs are, that we can now assume completeness and back-test tools to assess how they perform.These are minor issues can be addressed in updates.
000
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
CheckM2 still appeared to struggle with the orders SAR86 and GCA-002705445 (Gammaproteobacteria), most never reach >75% completeness, despite being circular. These groups should be treated with more care. CheckM2 performs much better than CheckM1, so should be the standard for marine metagenomics.
110
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
CheckM2 is much less problematic in this respect than CheckM1, which is no surprise because it’s the problem CheckM2 was meant to address, but highlights that past seawater MAG work using CheckM1 would likely have thrown out many SAR86, etc that should show higher completeness values than they do.
110
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
Because we recovered many circular MAGs we could assume to be complete, we could ask where CheckM1 and 2 work well and where they struggled. We assumed that all circular MAGs should show completeness values of 100%, and a handful of prok orders and classes were always underestimated (<85% comp).
110
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
In addition to GC and strain bias, certain taxa showed systematically underestimated completeness via CheckM1 and 2, so historically much more likely to get erroneously thrown away in fragmented short read MAGs—a 3rd reason for the existence of Ghost Taxa. A sub-🧵. bsky.app/profile/stev...
144
Reposted by Steven Robbins
Oxford Nanopore @nanoporetech.com · 22/07/2026
For 10+ years, short reads left the ocean's most dominant microbes unstudied. Now we can finally see them. Steven & the team used nanopore reads to assemble the ocean's most abundant "Ghost Taxa." The microbes, invisible for a decade, could also reflect the condition of the reef around them. #WYMM
032
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
Looks like the reads are on ENA, MAGs are on Zenodo: zenodo.org/records/2050... Will check if there's a plan to upload MAGs to ENA
zenodo.org
Great Barrier Reef Microbial Genomes Database (GBR-MGD)
This record contains: (i) prokaryote genomes binned from seawater metagenomes collected from 48 sites on the Great Barrier Reef (ii) Crassvirales genomes identified in metagenome assemblies (iii) Mami...
000
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
Hey Daan. GlobDB is the reason I thought of you--great resource. Hmmm, the MAGs should be in ENA and/or NCBI...but i'm not at UQ anymore and didn't handle the upload. Let me send you an email to link you to the people that did.
110
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
Short-read metagenomic sequencing cannot recover genomes from many abundant marine prokaryotes due to high strain heterogeneity and platform-specific GC bias (likely viruses, too), but Nanopore long reads can address this. A results 🧵on our recent paper in @nature.com . bsky.app/profile/stev...
1105
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
@archaeal.bsky.social, @lauraeme.bsky.social, @minderoo.bsky.social, @benjwoodcroft.bsky.social, @danielsprockett.bsky.social, @aaronpomerantz.bsky.social, @nanoporetech.com, @michiwagner4.bsky.social, @surtlab.bsky.social, @sewerynoz.bsky.social, @daanspeth.bsky.social, @bigdatabiology.bsky.social.
100
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
Pinging potentially interested parties :-) : @reefgenomics.bsky.social, @peixotoraquel.bsky.social, @merenbey.bsky.social, @jcamthrash.bsky.social, @kirk3gaard.bsky.social, @acritschristoph.bsky.social, @simrouxvirus.bsky.social, @alexjaf.bsky.social, @ace-uq.bsky.social, @ace-gtdb.bsky.social...
220
Steven Robbins @stevenjrobbins.bsky.social · 22/07/2026
It's out! Excited to present the Great Barrier Reef Microbial Genomes Database (GBR-MGD), a comprehensive DB of 1000s of high-quality prokaryote, virus, plasmid, and chromosome-level eukaryote MAGs using Nanopore long reads. Subthreads incoming. Please share widely. 🙂 www.nature.com/articles/s41...
nature.com
The planktonic microbiome of the Great Barrier Reef - Nature
The Great Barrier Reef Microbial Genomes Database compiles prokaryotic, viral and eukaryotic genomes from seawater collected from the Great Barrier Reef, providing a rich resource for the study of mar...
88443
Reposted by Steven Robbins
alex jaffe @alexjaf.bsky.social · 21/07/2026
my first 'data descriptor'! hoping this one will be useful for folks interested in tracing the distribution of genomes/genes/proteins over physicochemical gradients in the (deep) marine water column.
2178
Reposted by Steven Robbins
Dr. Jennifer Glass (she/her) @methanojen.bsky.social · 16/07/2026
We have a bunch of MAGs sitting in GenBank queues for many months. I run all MAGs through PGAP so they’re BLAST-able & assigned accessions, which I refer to when discussing genes in MAGs in papers. This has ground to a halt w/ NCBI processing delays in recent years. Is there anyway to expedite this?
443
Reposted by Steven Robbins
Evgenii Protasov @evgenii-protasov.bsky.social · 05/07/2026
Predicting oxygen levels in microbial habitats using a metagenome-based approach #microbiology #metagenomics #oxygen @asm.org doi.org/10.1128/msys...
doi.org
Predicting oxygen levels in microbial habitats using a metagenome-based approach | mSystems
Oxygen is one of the most important environmental variables affecting microbial activity and composition, but is often difficult to measure in situ. We developed a tool, OxyMetaG, that leverages differences in bacterial gene content across known aerobic and anaerobic taxa to predict the oxygen level of a given sample directly from shotgun metagenomic reads. OxyMetaG works on samples with low sequencing depth and avoids computationally expensive genome assembly, which often captures only a fraction of the microbial community in a given environment. With OxyMetaG, bacteria can be used as bioindicators of oxygen availability over broader time scales than just a single measurement and provide crucial environmental context in cases where oxygen has not been or cannot be measured. OxyMetaG is publicly available and can be used to answer a wide variety of ecological questions in both environmental and host-associated systems.
01110
Reposted by Steven Robbins
bioRxiv Genomics @biorxiv-genomic.bsky.social · 29/06/2026
Restriction-site-based enrichment coupled to adaptive sampling enables long-read transposon-insertion sequencing www.biorxiv.org/content/10.64898/20…
042
Reposted by Steven Robbins
Daan Speth @daanspeth.bsky.social · 26/06/2026
I'm happy to announce the release of GlobDB r232! This version contains 346,233 bacterial and archaeal genomes, based on 26 datasets. More info globdb.org 🦠🖥️🧬
globdb.org
home | GlobDB
24623