Alex Arzamasov

Some cool data on bifidobacteria and their relatives in honeybees!

Evolution of gut microbiota across honeybee species revealed by comparative metagenomics Nature Communications - Here, the authors perform metagenomic analysis of honeybee gut microbes, uncovering many previously unidentified species and host specific differences in composition and...

Happy to announce that our comparative metagenomic analysis of the gut microbiota of five honeybee species - spearheaded by @aiswarya.bsky.social - is finally published in a peer-reviewed journal! rdcu.be/eKMCs @fbm-unil.bsky.social @dmf-unil.bsky.social 🧵👇

Genomic studies in bifidobacteria often focus only on CAZymes. IMHO, in Gram-positives, transporters are the true gatekeepers of glycan metabolism. We need to prioritize improving their functional annotations and include them in metabolic reconstructions (7/7)

We demonstrated that phylogenetically closely related strains can exhibit substantial differences in HMO utilization, which is driven by subtle variations in HMO transporter genes. Species/subspecies names alone don’t tell the full story; one needs to look at gene content in individual strains(6/7)

Among the new pathways we uncovered was a xyloglucan degradation pathway that was present in rare B. catenulatum subsp. kashiwanohense strains and conserved in B. dentium and B. tsurumiense (5/7)

We validated phenotypic predictions for 30 bifidobacterial strains, achieving 94% accuracy. For example, we confirmed the unique ability of the new B. longum clade to grow on starch and pullulan, and described an unconventional B. adolescentis strain that can use 2’-fucosyllactose (4/7)

Our analysis revealed notable inter- and intra-species variability. Among notable findings was a new Bifidobacterium longum clade harboring pathways for starch, pullulan, and difructose dianhydride metabolism but lacking pathways for LNB/GNB, N-glycan, and human milk oligosaccharide utilization(3/7)

We reconstructed 68 glycan utilization pathways encoded in 3,083 bif genomes by looking at the distribution of 589 curated metabolic functions (transporters, CAZymes, etc). Several years of manual curation greatly improved the quality of functional gene annotations (>90% for transporters!) (2/7)

It’s been a bit over 2 months since the main results of my PhD work on reconstructing carbohydrate utilization pathways in human bifidobacteria were finally published; so I guess it’s a good time to update the previous thread (1/7) www.nature.com/articles/s41...

Among the new pathways we uncovered was a xyloglucan degradation pathway that was present in rare B. catenulatum subsp. kashiwanohense strains and conserved in B. dentium and B. tsurumiense (5/7)

We validated phenotypic predictions for 30 bifidobacterial strains, achieving 94% accuracy. For example, we confirmed the unique ability of the new B. longum clade to grow on starch and pullulan, and described a B. adolescentis strain that can use 2’-fucosyllactose (4/7)

Our analysis revealed notable inter- and intra-species variability. Among notable findings was a new Bifidobacterium longum clade harboring pathways for starch, pullulan, and difructose dianhydride metabolism but lacking pathways for LNB/GNB, N-glycan, and human milk oligosaccharide utilization(3/7)

We reconstructed 68 glycan utilization pathways encoded in 3,083 bif genomes by looking at the distribution of 589 curated metabolic functions (transporters, CAZymes, etc). Several years of manual curation greatly improved the quality of functional gene annotations (>90% for transporters!) (2/7)

Excited to share our @cp-cellhostmicrobe.bsky.social led by Magda showing how Bif has co-evolved with different animal hosts 🐒🐭🐷🐦

Key takeaways:
🔹 Host ancestry + diet shape Bif evolution
🔹 Mammals enriched for carb-busting enzymes
🔹 Untapped diversity in non-human hosts = new probiotic potential

Our study developing a skin metatranscriptomics protocol is now out in @natbiotech.nature.com!

We finally have the ability to study microbial activity on skin and identify key functional genes playing a role in diseases.

Amazing team of Chia Minghao and Amanda Ng 👏

nature.com/articles/s41...

Uptake of fucosylated type I human milk oligosaccharide blocks by Bifidobacterium longum subsp. infantis | mBio The assembly of the gut microbiota in early life is critical to the health trajectory of human hosts. Breast feeding selects for a Bifidobacterium-rich community, adapted to efficiently utilize human ...

Very nice paper that deciphers the transport mechanisms of fucosylated lacto-N-biose in Bifidobacterium longum subsp. infantis
journals.asm.org/doi/full/10....

High-throughput single-cell isolation of Bifidobacterium strains from the gut microbiome https://www.biorxiv.org/content/10.1101/2025.07.23.666462v1

Predicting the subcellular location of prokaryotic proteins with DeepLocPro AbstractMotivation. Protein subcellular location prediction is a widely explored task in bioinformatics because of its importance in proteomics research. W

Thanks! Also, maybe of interest, I think PSORTb and the new tool DeepLocPro perform better for bifido proteins than signalP
academic.oup.com/bioinformati...

Lacto-N-biosidase Encoded by a Novel Gene of Bifidobacterium longum Subspecies longum Shows Unique Substrate Specificity and Requires a Designated Chaperone for Its Active Expression * Infant gut-associated bifidobacteria possess species-specific enzymatic sets to assimilate human milk oligosaccharides, and lacto-N-biosidase (LNBase) is a key enzyme that degrades lacto-N-tetraose (G...

4) The GH136 lacto-N-biosidase from B. longum that you cite does not release LacNAc from Lacto-N-neotetraose (Fig. 4b from Sakurama et al). The B. bifidum GH136 is not identical to the B. longum one, but I haven't seen papers describing activity on LNnT (only LNT)
www.jbc.org/article/S002...

Structural and Thermodynamic Analyses of Solute-binding Protein from Bifidobacterium longum Specific for Core 1 Disaccharide and Lacto-N-biose I * Recently, a gene cluster involving a phosphorylase specific for lacto-N-biose I (LNB; Galβ1–3GlcNAc) and galacto-N-biose (GNB; Galβ1–3GalNAc) has been found in Bifidobacterium longum. We showed that t...

3) Additionally, B. bifidum possesses an ortholog of an ABC transporter that can uptake GNB (GltABC), so it would make sense for the GH101 enzyme to be extracellular, so the released disaccharide is then imported.
www.jbc.org/article/S002...

α-N-Acetylglucosaminidase from Bifidobacterium bifidum specifically hydrolyzes α-linked N-acetylglucosamine at nonreducing terminus of O-glycan on gastric mucin - Applied Microbiology and Biotechnolog... α-Linked N-acetylglucosamine is one of the major glyco-epitopes in O-glycan of gastroduodenal mucin. Here, we identified glycoside hydrolase (GH) family 89 α-N-acetylglucosaminidase, termed AgnB, from...

2) Your predictions about the localization of some other GHs (like GH89 and GH101) also differ from previously published literature. For example, the GH89 was described as a membrane-anchored extracellular enzyme before.
link.springer.com/article/10.1...

Identification and characterization of a sulfoglycosidase from Bifidobacterium bifidum implicated in mucin glycan utilization Abstract. Human gut symbiont bifidobacteria possess carbohydrate-degrading enzymes that act on the O-linked glycans of intestinal mucins to utilize those c

Hi, very cool to see this out! I have a couple of comments about the model. 1) I think the GH20 enzyme that removes GlcNAc-6S (BbhII) is extracellular. The released GlcNAc-6S can be cross-fed to Bifidobacterium breve (some strains have a dedicated utilization pathway)
doi.org/10.1080/0916...

Structural and Thermodynamic Analyses of Solute-binding Protein from Bifidobacterium longum Specific for Core 1 Disaccharide and Lacto-N-biose I * Recently, a gene cluster involving a phosphorylase specific for lacto-N-biose I (LNB; Galβ1–3GlcNAc) and galacto-N-biose (GNB; Galβ1–3GalNAc) has been found in Bifidobacterium longum. We showed that t...

3) Additionally, B. bifidum possesses an ortholog of an ABC transporter that can uptake GNB (GltABC), so it would make sense for this GH89 to be extracellular, so the released disaccharide is then imported.
www.jbc.org/article/S002...

PNGaseA-mediated N-glycan stripping from peptides by infant-derived Bifidobacterium bifidum N-glycans are highly common sources of nutrition for human colonic-dwelling bacteria. These microbes have evolved a several methods to remove N-glycans from proteins; herein we describe the biochemica...

New Pre-print! “PNGaseA-mediated N-glycan stripping from peptides by infant-derived Bifidobacterium bifidum”. This is the first manuscript from the Bifidobacterium and breast milk project🍼🤱👶in collaboration with van Sinderen and Lovering groups
www.biorxiv.org/content/10.1...

Is the original 1924 article publicly available somewhere?

skDER and CiDDER: two scalable approaches for microbial genome dereplication An abundance of microbial genomes have been sequenced in the past two decades. For fundamental comparative genomic investigations, where the goal is to determine the major gain and loss events shaping...

Great to see our manuscript on skDER & CiDDER - programs for selection of representative microbial genomes - now published.

Please give them a try and if you have any feature requests or issues, just let us know.

www.microbiologyresearch.org/content/jour...

github.com/raufs/skDER

UF, Trinity College team cracks micronutrient mystery that could be key to brain health, cancer defense Unlocking how our bodies absorb a micronutrient that we rely on for everything from healthy brain function to cancer defense.

The discovery of the missing Queuosine transporter is finally out. 15 years in the making with a combination of data mining and experimental validation performed by an international consortium funded by NIH and Irish agencies.
news.ufl.edu/2025/06/micr...

Functional gene annotation has been a problem in the comparative genomics of bacteria from the start, and now ML is only scaling the problem. IMHO, the real issue is that fixing annotations is often seen as too "incremental" to fund or publish. Why do it if it's easier to analyze a bigger dataset

Rachel Thomas, PhD - Deep learning gets the glory, deep fact checking gets ignored an AI researcher going back to school for immunology

A great blog post by @math-rachel.bsky.social on this preprint describing failures of an ML approach for gene annotation:
rachel.fast.ai/posts/2025-0...
www.biorxiv.org/content/10.1...

"how challenging (or even impossible) it can be to evaluate AI claims in work outside our own area of expertise"

I often flag genomic predictions of SCFA production that clash with “textbook” data. But sometimes it’s the experimental claims that raise eyebrows. This recent paper reports propionate (!) and butyrate (!!) production by Bif. longum ssp. infantis (it shouldn't make either
doi.org/10.3168/jds....