Aude Bernheim
@audeber.bsky.social
PI at Institut Pasteur
Evolution, immunity, genomics, microbiolgy.
Into immunity in bacteria and its conservation in eukaryotes.
Advocate for more inclusive sciences
https://research.pasteur.fr/en/team/molecular-diversity-of-microbes/
Evolution, immunity, genomics, microbiolgy.
Into immunity in bacteria and its conservation in eukaryotes.
Advocate for more inclusive sciences
https://research.pasteur.fr/en/team/molecular-diversity-of-microbes/
Overall we discovered a novel family of immune genes conserved across bacteria and eukaryotes.
SIRal, a human homolog, is a novel actor of the TLR pathway.
Much remains to be understood about SIRal and Sirims in immunity, but 🦠 will help us along the way.
SIRal, a human homolog, is a novel actor of the TLR pathway.
Much remains to be understood about SIRal and Sirims in immunity, but 🦠 will help us along the way.
July 24, 2025 at 6:23 PM
Overall we discovered a novel family of immune genes conserved across bacteria and eukaryotes.
SIRal, a human homolog, is a novel actor of the TLR pathway.
Much remains to be understood about SIRal and Sirims in immunity, but 🦠 will help us along the way.
SIRal, a human homolog, is a novel actor of the TLR pathway.
Much remains to be understood about SIRal and Sirims in immunity, but 🦠 will help us along the way.
How does this family of proteins function?
Like their bacterial relatives, eukaryotic SIRims — including human SIRal — use NAD+ in vitro. This enzymatic activity (NADase) seems ancient and conserved across the tree of life.
Like their bacterial relatives, eukaryotic SIRims — including human SIRal — use NAD+ in vitro. This enzymatic activity (NADase) seems ancient and conserved across the tree of life.
July 24, 2025 at 6:23 PM
How does this family of proteins function?
Like their bacterial relatives, eukaryotic SIRims — including human SIRal — use NAD+ in vitro. This enzymatic activity (NADase) seems ancient and conserved across the tree of life.
Like their bacterial relatives, eukaryotic SIRims — including human SIRal — use NAD+ in vitro. This enzymatic activity (NADase) seems ancient and conserved across the tree of life.
What about pathogens?
We demonstrate that SIRal limits infection by herpes virus and by the bacteria Salmonella in mouse macrophages.
We demonstrate that SIRal limits infection by herpes virus and by the bacteria Salmonella in mouse macrophages.
July 24, 2025 at 6:23 PM
What about pathogens?
We demonstrate that SIRal limits infection by herpes virus and by the bacteria Salmonella in mouse macrophages.
We demonstrate that SIRal limits infection by herpes virus and by the bacteria Salmonella in mouse macrophages.
We analyzed SIRal’s function in human and mouse cells.
Through numerous experiments (yay bar graphs!), we show that SIRal is required for a key immune pathway in animals (Toll-like receptor/TLR). SIRal mediates the up-regulation of interferon-stimulated genes (ISGs) and proinflammatory cytokines.
Through numerous experiments (yay bar graphs!), we show that SIRal is required for a key immune pathway in animals (Toll-like receptor/TLR). SIRal mediates the up-regulation of interferon-stimulated genes (ISGs) and proinflammatory cytokines.
July 24, 2025 at 6:23 PM
We analyzed SIRal’s function in human and mouse cells.
Through numerous experiments (yay bar graphs!), we show that SIRal is required for a key immune pathway in animals (Toll-like receptor/TLR). SIRal mediates the up-regulation of interferon-stimulated genes (ISGs) and proinflammatory cytokines.
Through numerous experiments (yay bar graphs!), we show that SIRal is required for a key immune pathway in animals (Toll-like receptor/TLR). SIRal mediates the up-regulation of interferon-stimulated genes (ISGs) and proinflammatory cytokines.
SIRim proteins are conserved across highly diverse phyla from bacteria to fish, fungi and protists.🐟🍄🐘
Among this large family of proteins, present in thousands of living organisms, we identified on human protein, which function was previously unknown.
We named it SIRal, for SIR2 antiphage-like.
Among this large family of proteins, present in thousands of living organisms, we identified on human protein, which function was previously unknown.
We named it SIRal, for SIR2 antiphage-like.
July 24, 2025 at 6:23 PM
SIRim proteins are conserved across highly diverse phyla from bacteria to fish, fungi and protists.🐟🍄🐘
Among this large family of proteins, present in thousands of living organisms, we identified on human protein, which function was previously unknown.
We named it SIRal, for SIR2 antiphage-like.
Among this large family of proteins, present in thousands of living organisms, we identified on human protein, which function was previously unknown.
We named it SIRal, for SIR2 antiphage-like.
What if this immune function also exists in humans, but hadn’t been discovered yet?
We analyzed SIR2 from thousands of bacterial, archaeal & eukaryotic genomes. All housekeeping sirtuins group in a clade, whereas all antiphage SIR2 group in another one that we named SIRim, for “SIR” and “immunity”.
We analyzed SIR2 from thousands of bacterial, archaeal & eukaryotic genomes. All housekeeping sirtuins group in a clade, whereas all antiphage SIR2 group in another one that we named SIRim, for “SIR” and “immunity”.
July 24, 2025 at 6:23 PM
What if this immune function also exists in humans, but hadn’t been discovered yet?
We analyzed SIR2 from thousands of bacterial, archaeal & eukaryotic genomes. All housekeeping sirtuins group in a clade, whereas all antiphage SIR2 group in another one that we named SIRim, for “SIR” and “immunity”.
We analyzed SIR2 from thousands of bacterial, archaeal & eukaryotic genomes. All housekeeping sirtuins group in a clade, whereas all antiphage SIR2 group in another one that we named SIRim, for “SIR” and “immunity”.
So great to see the work of the lab being brillantly presented at conferences across 🇮🇹🍦🍝, @matthieu-haudiquet.bsky.social on Evolution and Mechanism of Lamassu in FEMS, Milan and Ernest Mordret on LLM for prediction of antiphage systems in GRC Evolutionary Genomics,Tuscany.
Check out their work 👇
Check out their work 👇
July 17, 2025 at 5:39 PM
So great to see the work of the lab being brillantly presented at conferences across 🇮🇹🍦🍝, @matthieu-haudiquet.bsky.social on Evolution and Mechanism of Lamassu in FEMS, Milan and Ernest Mordret on LLM for prediction of antiphage systems in GRC Evolutionary Genomics,Tuscany.
Check out their work 👇
Check out their work 👇
Happy to represent bacterial immunity with great bacterial immunologist Tana Wein at a conference honouring the 40th birthday of Toll discovery, made in Drosophila.
Really fun to be celebrating the contribution of diverse organisms to our understanding of immunity 🪰🐘🦠.
Really fun to be celebrating the contribution of diverse organisms to our understanding of immunity 🪰🐘🦠.
June 10, 2025 at 1:12 PM
Happy to represent bacterial immunity with great bacterial immunologist Tana Wein at a conference honouring the 40th birthday of Toll discovery, made in Drosophila.
Really fun to be celebrating the contribution of diverse organisms to our understanding of immunity 🪰🐘🦠.
Really fun to be celebrating the contribution of diverse organisms to our understanding of immunity 🪰🐘🦠.
It was an honor for my lab and myself to host the meeting on the Immune Systems of Bacteria. Such a high from all the incredible science and sharing with scientists from the whole world the greatness of Paris in the spring.
Vive la science, et vive Paris!
Vive la science, et vive Paris!
April 11, 2025 at 3:46 PM
It was an honor for my lab and myself to host the meeting on the Immune Systems of Bacteria. Such a high from all the incredible science and sharing with scientists from the whole world the greatness of Paris in the spring.
Vive la science, et vive Paris!
Vive la science, et vive Paris!
The EMBO workshop on the Immune System of Bacteria just started in sunny Paris in the Institut Pasteur.
So excited for three days of awesome science.
So excited for three days of awesome science.
April 8, 2025 at 10:24 AM
The EMBO workshop on the Immune System of Bacteria just started in sunny Paris in the Institut Pasteur.
So excited for three days of awesome science.
So excited for three days of awesome science.
We then move on to a database of >30k diverse prok. genomes to chart a landscape of predicted antiphage proteins.
We built an interactive UMAP to geek this space mdmparis.github.io/antiphage-la...
Quantitatively, rarefaction curves led us to estimate that probably >45k prot families are antiphage
We built an interactive UMAP to geek this space mdmparis.github.io/antiphage-la...
Quantitatively, rarefaction curves led us to estimate that probably >45k prot families are antiphage
January 9, 2025 at 1:18 PM
We then move on to a database of >30k diverse prok. genomes to chart a landscape of predicted antiphage proteins.
We built an interactive UMAP to geek this space mdmparis.github.io/antiphage-la...
Quantitatively, rarefaction curves led us to estimate that probably >45k prot families are antiphage
We built an interactive UMAP to geek this space mdmparis.github.io/antiphage-la...
Quantitatively, rarefaction curves led us to estimate that probably >45k prot families are antiphage
These systems, named after soil deities (as Streptomyces are mostly soil bacteria) are super cool!
Geb is only 150AA protein with no annotated Pfam domain while Ukko is a 4 genes system with intriguing links to metabolism.
Going beyond classic model organisms is awesome for novel biology.
Geb is only 150AA protein with no annotated Pfam domain while Ukko is a 4 genes system with intriguing links to metabolism.
Going beyond classic model organisms is awesome for novel biology.
January 9, 2025 at 1:18 PM
These systems, named after soil deities (as Streptomyces are mostly soil bacteria) are super cool!
Geb is only 150AA protein with no annotated Pfam domain while Ukko is a 4 genes system with intriguing links to metabolism.
Going beyond classic model organisms is awesome for novel biology.
Geb is only 150AA protein with no annotated Pfam domain while Ukko is a 4 genes system with intriguing links to metabolism.
Going beyond classic model organisms is awesome for novel biology.
Next, we validated these predictions and checked if we could find really "novel" things.
By novel, we mean 1)not picked up by defense score, and 2) with proteins with domains never involved in antiphage defense before.
We validated six novel antiphage systems in Streptomyces (out of 10 tested)!
By novel, we mean 1)not picked up by defense score, and 2) with proteins with domains never involved in antiphage defense before.
We validated six novel antiphage systems in Streptomyces (out of 10 tested)!
January 9, 2025 at 1:18 PM
Next, we validated these predictions and checked if we could find really "novel" things.
By novel, we mean 1)not picked up by defense score, and 2) with proteins with domains never involved in antiphage defense before.
We validated six novel antiphage systems in Streptomyces (out of 10 tested)!
By novel, we mean 1)not picked up by defense score, and 2) with proteins with domains never involved in antiphage defense before.
We validated six novel antiphage systems in Streptomyces (out of 10 tested)!
We benchmarked these algorithms against each other & the "classic" defense score (enrichment in defense islands) and found that these approaches are complementary !
Some systems found in hotspots of MGE could be missed by defense score, and some novel systems without known antiphage domains by ESM.
Some systems found in hotspots of MGE could be missed by defense score, and some novel systems without known antiphage domains by ESM.
January 9, 2025 at 1:18 PM
We benchmarked these algorithms against each other & the "classic" defense score (enrichment in defense islands) and found that these approaches are complementary !
Some systems found in hotspots of MGE could be missed by defense score, and some novel systems without known antiphage domains by ESM.
Some systems found in hotspots of MGE could be missed by defense score, and some novel systems without known antiphage domains by ESM.
Then, genomic language model. In this framework, the vocabulary corresponds to protein families. We trained a language model to “read” stretches of proteins on bacterial genomes.
We focused on Actinobacteria that have been underexplored for their antiphage repertoire.
👋ALBERT-DefenseFinder🧐
We focused on Actinobacteria that have been underexplored for their antiphage repertoire.
👋ALBERT-DefenseFinder🧐
January 9, 2025 at 1:18 PM
Then, genomic language model. In this framework, the vocabulary corresponds to protein families. We trained a language model to “read” stretches of proteins on bacterial genomes.
We focused on Actinobacteria that have been underexplored for their antiphage repertoire.
👋ALBERT-DefenseFinder🧐
We focused on Actinobacteria that have been underexplored for their antiphage repertoire.
👋ALBERT-DefenseFinder🧐
First, we fined-tuned a protein language model (where protein sequences are the input) on known antiphage proteins and called it ESM-DefenseFinder.
We show ESM-DefenseFinder can predict many novel antiphage proteins, which represent variants of known systems as it relies mostly on homology.
We show ESM-DefenseFinder can predict many novel antiphage proteins, which represent variants of known systems as it relies mostly on homology.
January 9, 2025 at 1:18 PM
First, we fined-tuned a protein language model (where protein sequences are the input) on known antiphage proteins and called it ESM-DefenseFinder.
We show ESM-DefenseFinder can predict many novel antiphage proteins, which represent variants of known systems as it relies mostly on homology.
We show ESM-DefenseFinder can predict many novel antiphage proteins, which represent variants of known systems as it relies mostly on homology.
In recent years, > 200 novel antiphage systems were discovered and characterized. DefenseFinder defensefinder.mdmlab.fr, allows to map the currently known diversity.
But computational approaches like defense islands, MGE hotspots suggest thousands remain undescribed. How big is this unknown?
But computational approaches like defense islands, MGE hotspots suggest thousands remain undescribed. How big is this unknown?
January 9, 2025 at 1:18 PM
In recent years, > 200 novel antiphage systems were discovered and characterized. DefenseFinder defensefinder.mdmlab.fr, allows to map the currently known diversity.
But computational approaches like defense islands, MGE hotspots suggest thousands remain undescribed. How big is this unknown?
But computational approaches like defense islands, MGE hotspots suggest thousands remain undescribed. How big is this unknown?
What is the diversity of antiphage systems out there ?
We used protein and genomic language models (and defense score!) to start bringing some answers: >45 000 protein families.
Leb by E. Mordret.
www.biorxiv.org/content/10.1...
Interactive UMAP to have fun
mdmparis.github.io/antiphage-la...
We used protein and genomic language models (and defense score!) to start bringing some answers: >45 000 protein families.
Leb by E. Mordret.
www.biorxiv.org/content/10.1...
Interactive UMAP to have fun
mdmparis.github.io/antiphage-la...
January 9, 2025 at 1:18 PM
What is the diversity of antiphage systems out there ?
We used protein and genomic language models (and defense score!) to start bringing some answers: >45 000 protein families.
Leb by E. Mordret.
www.biorxiv.org/content/10.1...
Interactive UMAP to have fun
mdmparis.github.io/antiphage-la...
We used protein and genomic language models (and defense score!) to start bringing some answers: >45 000 protein families.
Leb by E. Mordret.
www.biorxiv.org/content/10.1...
Interactive UMAP to have fun
mdmparis.github.io/antiphage-la...
Registration now open for the EMBO Workshop: Immune System of Bacteria !!!!
www.sisb2025.conferences-pasteur.org/home
Joins us in Institut Pasteur, Paris, April 8-10th.
www.sisb2025.conferences-pasteur.org/home
Joins us in Institut Pasteur, Paris, April 8-10th.
December 9, 2024 at 12:52 PM
Registration now open for the EMBO Workshop: Immune System of Bacteria !!!!
www.sisb2025.conferences-pasteur.org/home
Joins us in Institut Pasteur, Paris, April 8-10th.
www.sisb2025.conferences-pasteur.org/home
Joins us in Institut Pasteur, Paris, April 8-10th.
Today, I had the honor of being awarded the Irene-Joliot Curie Prize (young woman scientist) at the French Academy of Science.
Such a beautiful place, moved and happy ++ to share it with Hugo, Florian and Jean who started the lab with me, and in presence of scientists who inspire me !
Such a beautiful place, moved and happy ++ to share it with Hugo, Florian and Jean who started the lab with me, and in presence of scientists who inspire me !
November 26, 2024 at 9:27 PM
Today, I had the honor of being awarded the Irene-Joliot Curie Prize (young woman scientist) at the French Academy of Science.
Such a beautiful place, moved and happy ++ to share it with Hugo, Florian and Jean who started the lab with me, and in presence of scientists who inspire me !
Such a beautiful place, moved and happy ++ to share it with Hugo, Florian and Jean who started the lab with me, and in presence of scientists who inspire me !
Join us in Paris on April 8th-10th for the 2025 Symposium on the Immune System of Bacteria 🥖🍷🦠🧬.
Excited to welcome these amazing speakers at Institut Pasteur.
Registration will open beginning of November. Abstract deadline February 28. Stay tuned.
Co-organized by @soreklab.bsky.social P. Kranzusch
Excited to welcome these amazing speakers at Institut Pasteur.
Registration will open beginning of November. Abstract deadline February 28. Stay tuned.
Co-organized by @soreklab.bsky.social P. Kranzusch
October 21, 2024 at 2:32 PM
Join us in Paris on April 8th-10th for the 2025 Symposium on the Immune System of Bacteria 🥖🍷🦠🧬.
Excited to welcome these amazing speakers at Institut Pasteur.
Registration will open beginning of November. Abstract deadline February 28. Stay tuned.
Co-organized by @soreklab.bsky.social P. Kranzusch
Excited to welcome these amazing speakers at Institut Pasteur.
Registration will open beginning of November. Abstract deadline February 28. Stay tuned.
Co-organized by @soreklab.bsky.social P. Kranzusch
SIRanc is 1 out of 5 independent clades of sirim proteins in eukaryotes, suggesting at least 5 horizontal gene transfer events from bacteria to eukaryotes.
In total 19% of queried eukaryotic genomes encode a sirim, spanning 189 very diverse species. So much to explore 🤩!
In total 19% of queried eukaryotic genomes encode a sirim, spanning 189 very diverse species. So much to explore 🤩!
September 23, 2024 at 9:35 AM
SIRanc is 1 out of 5 independent clades of sirim proteins in eukaryotes, suggesting at least 5 horizontal gene transfer events from bacteria to eukaryotes.
In total 19% of queried eukaryotic genomes encode a sirim, spanning 189 very diverse species. So much to explore 🤩!
In total 19% of queried eukaryotic genomes encode a sirim, spanning 189 very diverse species. So much to explore 🤩!
So SIRanc is involved in mammalian immunity, but what is the mechanism?
Bacterial SIRim degrade NAD+, could SIRanc also degrade NAD+? Indeed! SIRanc has NADase activity in vitro, and this activity is necessary for SIRanc in vivo.
So SIRanc like bacterial SIRim is an NADase!
Bacterial SIRim degrade NAD+, could SIRanc also degrade NAD+? Indeed! SIRanc has NADase activity in vitro, and this activity is necessary for SIRanc in vivo.
So SIRanc like bacterial SIRim is an NADase!
September 23, 2024 at 9:34 AM
So SIRanc is involved in mammalian immunity, but what is the mechanism?
Bacterial SIRim degrade NAD+, could SIRanc also degrade NAD+? Indeed! SIRanc has NADase activity in vitro, and this activity is necessary for SIRanc in vivo.
So SIRanc like bacterial SIRim is an NADase!
Bacterial SIRim degrade NAD+, could SIRanc also degrade NAD+? Indeed! SIRanc has NADase activity in vitro, and this activity is necessary for SIRanc in vivo.
So SIRanc like bacterial SIRim is an NADase!
We took primary human monocyte-derived macrophages (huMDMs) and knocked them down (KD) for SIRanc expression. We monitored mRNA specific of the Myd88 or the TRIF-dependent branch of the TLR4 pathway. We observed the same thing as in mice!
=> SIRanc contributes to TLR signalling!
=> SIRanc contributes to TLR signalling!
September 23, 2024 at 9:33 AM
We took primary human monocyte-derived macrophages (huMDMs) and knocked them down (KD) for SIRanc expression. We monitored mRNA specific of the Myd88 or the TRIF-dependent branch of the TLR4 pathway. We observed the same thing as in mice!
=> SIRanc contributes to TLR signalling!
=> SIRanc contributes to TLR signalling!
Cool, but in which branch of the TLR4 pathway?
Through a series of experiments, we show that SIRanc is involved in the TRIF dependent (and not MyD88) branch of the TLR4 pathway in mice, a reason why it could have been missed by other screens !
But what about humans?
Through a series of experiments, we show that SIRanc is involved in the TRIF dependent (and not MyD88) branch of the TLR4 pathway in mice, a reason why it could have been missed by other screens !
But what about humans?
September 23, 2024 at 9:33 AM
Cool, but in which branch of the TLR4 pathway?
Through a series of experiments, we show that SIRanc is involved in the TRIF dependent (and not MyD88) branch of the TLR4 pathway in mice, a reason why it could have been missed by other screens !
But what about humans?
Through a series of experiments, we show that SIRanc is involved in the TRIF dependent (and not MyD88) branch of the TLR4 pathway in mice, a reason why it could have been missed by other screens !
But what about humans?