Tag: Papers

Ákos Kovács had the brilliant idea of putting up a temporary resource for things you bring up in a talk that you can point people to. I did not do this before my talk at ISME today, but I thought the idea was so good, so here’s a summary and collection of my ISME short-talk on the EMBARK outcomes today:

• More information on EMBARK and its successor SEARCHER can be found on the project website, here: http://antimicrobialresistance.eu Importantly, this is a team effort over four years and I only touched on a few selected things
• Within the project we have looked at typical background levels of antibiotic resistance in the environment. We have already published some of these results (for qPCR abundances) in Abramova et al. 2023
• The average resistance gene in the average environment is present in ~1 in 1000 bacteria, but the variation between different genes is huge
• Depending on monitoring goal, different target genes are relevant to use. See this table adapted from Abramova et al. 2023:

• We have also tried to make different monitoring methods for environmental AMR comparable. Those mentioned in the talk were selective culturing for resistant bacteria, qPCR and shotgun metagenomics
• This data is not yet published, but overall we see relatively good correlation between qPCR and metagenomics. This is not true for all genes, though, and unfortunately neither qPCR nor metagenomics is always better than the other
• Culturing data is not very good at predicting specific antibiotic resistance gene abundances as the class level
• Finally, we have developed methods for discovering new types of ARGs, as seen in the ResFinderFG database: Gschwind et al. 2023
• We have also used these new methods to look at differences between established ARGs and latent ARGs in a variety of environments: Inda-Díaz et al. 2023
• Our ultimate goal in EMBARK would be to develop a modular framework for environmental monitoring of antibiotic resistance. You can read more about our thinking and goals in the review paper we published last year: Bengtsson-Palme et al. 2023

In just a few days, Environment International has published two papers coming out from the EMBARK consortium which are somewhat connected to each other.

The first (or technically the second, but the other order makes more sense when explaining this…) is the first paper involving most of the people who have been working in the EMBARK consortium for an extended period of time. It’s an overview paper titled “Towards monitoring of antimicrobial resistance in the environment: For what reasons, how to implement Itit, and what are the data needs?” (1) and I think the title describes the topic pretty well. Basically, we go through why it would be interesting to monitor for antibiotic resistance in the environments, how that could be implemented and what we would need to know to get there.

The very condensed story is that if one is considering implementing monitoring for environmental resistance, these are a few things that should be considered:

• The purpose of monitoring: What is the motivation? What should be achieved? What type of risk should be assessed? What type of action would monitoring enable?
• Choice of methods: Which methods are economically feasible? Which methods would deliver results within a useful timeframe for taking appropriate actions?
• Targeted environments: In what type of environment would monitoring for a given purpose be worthwhile?
• Intended users: Who would be able to use, implement and act upon this strategy?
• Integration potential: How does this monitoring integrate with other monitoring efforts? How can the resulting data be communicated?

We then dive into the knowledge gaps we are currently facing, and particularly highlight the following areas:

• Establish how different existing methods for monitoring resistance compare to each other
• Extend pathogen-centric databases for resistance genes with latent resistance genes (2)
• Determine the locations and type of environments relevant for resistance monitoring
  To reduce costs, utilizing already existing environmental monitoring should be prioritized, as should locations integrated into operating or planned surveillance programs. More efforts should also be made to identify additional pathways for resistance transmission through the environment.
• Study the environment as a source and transmission route for antibiotic resistance
  Stratify risks associated with resistance genes found in the environment. Define typical levels of antibiotic resistance in different environments (3), and define how these levels change over time.
• Identify settings where the relationship between fecal bacteria and antibiotic resistance is absent
  Usually, these levels follow each other, but the environments where they don’t are important as they deviate from the expected baseline of resistance. This knowledge can aid in identifying situations in which it would be helpful to investigate a microbial community for resistance to specific antibiotics.
• Identify the origins for more antibiotic resistance genes (4)
  This knowledge will be instrumental in preventing the emergence of new forms of resistance in pathogens in the future.

An important outcome of this paper is that we realise that we are still not at a level of understanding where routine monitoring for resistance in the environment can be easily justified or implemented. Still, there is a need for monitoring data in natural environments to even get started, and therefore we support the implementation of national, regional and global of initiatives without having all the scientific answers. The lack of comprehensive understanding should not be an obstacle to starting environmental monitoring for AMR, nor for action against environmental development and spread of AMR.

The second paper is very much related to the first, in that it actually tries to address one of these knowledge gaps: the need for normal background levels of antibiotic resistance in different environments. In this paper, Anna Abramova did an herculean effort collecting (we hope) all qPCR data on antibiotic resistance gene abundances in the environment for the past two decades. All in all, she collected data for more than 1500 samples across 150 studies and integrated these into an analys of what we could consider normal levels of resistance in different environments.

For an ‘average’ resistance gene, we found that the normal relative abundance range was form 10^-5 to 10^-3 copies per bacterial 16S rRNA, or that around one in 1,000 bacteria would carry a given resistance gene. This level varied quite a bit between different resistance genes, however, but not so much between environmental types (except for in human and animal feces, where some resistance genes were clearly more abundant, most prominently tetracycline resistance genes). What was more striking was that there was a clear difference between environments impacted or likely impacted by human activities, as opposed to more pristine environments with little to none human impact. Some resistance genes, such as tetA, tetG, blaTEM and blaCTX-M, showed very marked differences between these impacted and non-impacted environments, making them great markers of human-activity-associated resistance.

Our final recommendations with regards to monitoring include:

• Include the intI1, sul1, blaTEM, blaCTX-M and qnrS genes in environmental monitoring, along with a selection of tetracycline resistance genes, including either tetA or tetG.
• Other potential target genes could be sul3, vanA, tetH, aadA2, floR, ereA and mexF, which are abundant in some environments, but are not often included in qPCR studies of environmental AMR
• If a gene deviates from the expected 10^-5 to 10^-3 interval, this warrants further investigation of the causes.
• Maximum acceptable levels need to be determined not only taking relative abundances of genes into account, but also risks to human health as well as the numbers of bacteria in a given volume of sample into account (5,6) and transmission routes to humans (7)
• The different standards of reporting DNA abundances constituted a complicating factor for this study. Both abundances of resistance genes relative to the 16S rRNA gene and to the sample volume or weight should be reported.
• The absence of clear trends of increases or decreases in resistance gene abundances over time indicates a need for more systematic time series data in a variety of environments.

Our results also highlighted the scarcity of resistance gene data from parts of the world, particularly from Africa and South America, and underscores the need for a concerted effort to quantify typical background levels of resistance in the environment more broadly to enable efficient environmental surveillance schemes akin to those that exist in clinical and veterinary settings.

I encourage anyone with an interesting these topics to at least skim the full papers [Monitoring overview paper here, Normal qPCR resistance abundances here]. These will be great resources and I am very proud of them both. I would really like to thank the entire EMBARK team and our collaborators in CORNELIA, WastPAN and in other organisations. I would also like to thank Anna for her hard work on collecting and analysing the qPCR data for around two years. It has been a long ride, and I think we are both happy, proud and a bit relieved to finally see this paper published!

References

1. Bengtsson-Palme J, Abramova A, Berendonk TU, Coelho LP, Forslund SK, Gschwind R, Heikinheimo A, Jarquin-Diaz VH, Khan AA, Klümper U, Löber U, Nekoro M, Osińska AD, Ugarcina Perovic S, Pitkänen T, Rødland EK, Ruppé E, Wasteson Y, Wester AL, Zahra R: Towards monitoring of antimicrobial resistance in the environment: For what reasons, how to implement it, and what are the data needs? Environment International, 108089 (2023). doi: 10.1016/j.envint.2023.108089
2. Inda-Díaz JS, Lund D, Parras-Moltó M, Johnning A, Bengtsson-Palme J, Kristiansson E: Latent antibiotic resistance genes are abundant, diverse, and mobile in human, animal, and environmental microbiomes. Microbiome, 11, 44 (2023). doi: 10.1186/s40168-023-01479-0
3. Abramova A, Berendonk TU, Bengtsson-Palme J: A global baseline for qPCR-determined antimicrobial resistance gene prevalence across environments. Environment International, 178, 108084 (2023). doi: 10.1016/j.envint.2023.108084
4. Ebmeyer S, Kristiansson E, Larsson DGJ: A framework for identifying the recent origins of mobile antibiotic resistance genes. Communications Biology, 4 (2021). doi:10.1038/s42003-020-01545-5
5. Larsson DGJ, Andremont A, Bengtsson-Palme J, Brandt KK, de Roda Husman AM, Fagerstedt P, Fick J, Flach C-F, Gaze WH, Kuroda M, Kvint K, Laxminarayan R, Manaia CM, Nielsen KM, Ploy M-C, Segovia C, Simonet P, Smalla K, Snape J, Topp E, van Hengel A, Verner-Jeffreys DW, Virta MPJ, Wellington EM, Wernersson A-S: Critical knowledge gaps and research needs related to the environmental dimensions of antibiotic resistance. Environment International, 117, 132–138 (2018). doi: 10.1016/j.envint.2018.04.041
6. Pruden A, Larsson DGJ, Amézquita A, Collignon P, Brandt KK, Graham DW, et al. Management options for reducing the release of antibiotics and antibiotic resistance genes to the environment. Environmental Health Perspectives, 121, 878–885 (2013). doi:10.1289/ehp.1206446
7. Bengtsson-Palme J, Kristiansson E, Larsson DGJ: Environmental factors influencing the development and spread of antibiotic resistance. FEMS Microbiology Reviews, 42, 1, 68–80 (2018). doi: 10.1093/femsre/fux053

Together with our collaborators in Tromsø in Norway, we published a paper over the weekend in eBioMedicine describing the early colonization patterns of preterm infants, both in terms of the microbes that arrive early to the infants, but also in terms of the antibiotic resistance genes they carry.

In the paper (1), which is a continuation of an earlier study by part of the team (2), we analysed metagenomic data from six Norwegian neonatal intensive care units to better understand the bacterial microbiota of infants born preterm or on term and receiving different treatments. These groups included probiotic-supplemented and antibiotic-exposed extremely preterm infants (n = 29), antibiotic-exposed very preterm infants (n = 25), antibiotic-unexposed very preterm infants (n = 8), and antibiotic-unexposed full-term infants (n = 10). Stool samples were collected from the infants after 7, 28, 120, and 365 days of life and were analysed using shotgun metagenomics. We were particularly interested in the maturation of the preterm infant microbiome into a ‘normal’ healthy gut microbiome, and the colonization with bacteria carrying antibiotic resistance genes.

We found that microbiota maturation was largely determined by the length of hospitalisation for the infants and how much preterm they were. The use of probiotics rendered the gut microbiota and resistome of extremely preterm infants more alike to term infants on day 7 and partially restored the loss of species interconnectivity and stability associated with preterm delivery. Finally, colonisation with Escherichia coli was associated with the highest number of antibiotic-resistance genes in the infant microbiomes, followed by Klebsiella pneumoniae and Klebsiella aerogenes.

Being born very preterm, along with prolonged hospitalisation and frequent antibiotic use alters early life resistome and mobilome, leading to an increased gut carriage of antibiotic resistance genes and mobile genetic elements. On the other hand, the effect of probiotics was not unidirectional. Probiotics decreased resistome burden, but at the same time the bacterial strains in the probiotics appear to promote the activity of mobile genetic elements. Here, further study of the gut microbiota is necessary to be able to design strategies aiming to lower disease risk in vulnerable preterm infants.

As mentioned, this study was a collaboration with Veronika Pettersen‘s group in Tromsø, particularly Ahmed Bargheet, who have done a fabulous job on the bioinformatics and analysis of this study. I hope that we will continue this collaboration in the future (first step will be me visting Tromsø again in June!) This also continues a nice little “sidetrack” of the group’s research into the early life microbiome – previously represented by the work of Katariina Pärnänen (3) and Tove Wikström‘s vaginal microbiome study (4), which is a very interesting and relevant subject in terms of both medicine and microbial ecology. We are also setting up new collaborations in this area, so I hope that more will come out of this track in the next couple of years.

Finally, thank you Veronika for inviting me to participate in this great project!

References

1. Bargheet A, Klingenberg C, Esaiassen E, Hjerde E, Cavanagh JP, Bengtsson-Palme J, Pettersen VK: Development of early life gut resistome and mobilome across gestational ages and microbiota-modifying treatments. eBio Medicine, 92, 104613 (2023). doi: 10.1016/j.ebiom.2023.104613
2. Esaiassen E, Hjerde E, Cavanagh JP, Pedersen T, Andresen JH, Rettedal SI, Støen R, Nakstad B, Willassen NP, Klingenberg C: Effects of Probiotic Supplementation on the Gut Microbiota and Antibiotic Resistome Development in Preterm Infants. Frontiers in Pediatrics, 16, 6, 347 (2018). doi: 10.3389/fped.2018.00347
3. Pärnänen K, Karkman A, Hultman J, Lyra C, Bengtsson-Palme J, Larsson DGJ, Rautava S, Isolauri E, Salminen S, Kumar H, Satokari R, Virta M: Maternal gut and breast milk microbiota affect infant gut antibiotic resistome and mobile genetic elements. Nature Communications, 9, 3891 (2018). doi: 10.1038/s41467-018-06393-w
4. Wikström T, Abrahamsson S, Bengtsson-Palme J, Ek CJ, Kuusela P, Rekabdar E, Lindgren P, Wennerholm UB, Jacobsson B, Valentin L, Hagberg H: Microbial and human transcriptome in vaginal fluid at midgestation: association with spontaneous preterm delivery. Clinical and Translational Medicine, 12, 9, e1023 (2022). doi: 10.1002/ctm2.1023

What is the latent resistome? This is a term we coin in a new paper published yesterday in Microbiome. In the paper, we distinguish between the small number antibiotic resistance genes (ARGs) that are established, well-characterized, and available in existing resistance gene databases (what we refer to as “established ARGs”). These are typically ARGs encountered in clinical pathogens and are often already causing problems in human and animal infections. The remaining latently present ARGs, which we denote “latent ARGs”, are less or not at all studied, and are therefore much harder to detect (1). These latent ARGs are typically unknown and generally overlooked in most studies of resistance. They are also seldom accounted for in risk assessments of antibiotic resistance (2-4). This means that our view of the resistome and its diversity is incomplete, which hampers our ability to assess risk for promotion and spread of yet undiscovered resistance determinants.

In our new study, we try to alleviate this issue by analyzing more than 10,000 metagenomic samples. We show that the latent ARGs are more abundant and diverse than established ARGs in all studied environments, including the human- and animal-associated microbiomes. The total pan-resistomes, i.e., all ARGs present in an environment (including the latent ARGs), are heavily dominated by these latent ARGs. In contrast, the core resistome (the ARGs that are commonly encountered) comprise both latent and established ARGs.

In the study, we identified several latent ARGs that were shared between environments or that are already present in human pathogens. These are often located on mobile genetic elements that can be transferred between bacteria. Finally, we also show that wastewater microbiomes have surprisingly large pan- and core-resistomes, which makes this environment a potent high-risk environment for mobilization and promotion of latent ARGs, which may make it into pathogens in the future.

It is also interesting to note that this new study echoes the results of my own study from 2018, showing that soil and water environments contain a high diversity of latent ARGs (or ARGs not found in pathogens as I put it in the 2018 study), despite being almost devoid of established ARGs (5).

This project has been a collaboration with Erik Kristiansson’s research group, and particularly with Juan Inda-Diáz. It has been great fun to work with them and I hope that we will keep this collaboration going into the future! The study can be read in its entirety here.

References

1. Inda-Díaz JS, Lund D, Parras-Moltó M, Johnning A, Bengtsson-Palme J, Kristiansson E: Latent antibiotic resistance genes are abundant, diverse, and mobile in human, animal, and environmental microbiomes. Microbiome, 11, 44 (2023). doi: 10.1186/s40168-023-01479-0 [Paper link]
2. Martinez JL, Coque TM, Baquero F: What is a resistance gene? Ranking risk in resistomes. Nature Reviews Microbiology 2015, 13:116–123. doi:10.1038/nrmicro3399
3. Bengtsson-Palme J, Larsson DGJ: Antibiotic resistance genes in the environment: prioritizing risks. Nature Reviews Microbiology, 13, 369 (2015). doi: 10.1038/nrmicro3399-c1
4. Bengtsson-Palme J: Assessment and management of risks associated with antibiotic resistance in the environment. In: Roig B, Weiss K, Thoreau V (Eds.) Management of Emerging Public Health Issues and Risks: Multidisciplinary Approaches to the Changing Environment, 243–263. Elsevier, UK (2019). doi: 10.1016/B978-0-12-813290-6.00010-X
5. Bengtsson-Palme J: The diversity of uncharacterized antibiotic resistance genes can be predicted from known gene variants – but not always. Microbiome, 6, 125 (2018). doi: 10.1186/s40168-018-0508-2

Last week, we published a paper which has been cooking for a long time. It is the result of years of hard work from particularly the first author – Tove Wikström – but also Sanna who did the bulk of the bioinformatic analysis with some help from me (well, I mostly contributed as a sounding board for ideas, but hopefully that was useful). The paper describes the gene expression of both the human host and the microbial community in the vagina during pregnancy and how the expressed genes (and the composition of bacteria) are linked to early births (1) and was published in Clinical and Translational Medicine.

We found 17 human genes potentially influencing preterm births. Most prominently the kallikrein genes (KLK2 and KLK3) and four different forms of of metallothioneins (MT1s) were higher in the preterm group than among fullterm women. These genes may be involved in inflammatory pathways associated with preterm birth.

We also found 11 bacterial species associated with preterm birth, but most of them had low occurrence and abundance. In contrary to some earlier studies, we saw no differences in bacterial diversity or richness between women who delivered preterm and women who delivered at term. Nor did Lactobacillus crispatus – often proposed to be protective against preterm birth (2,3) – seem to be a protective factor against preterm birth. However, most other studies have used DNA-based approaches to determine the bacterial community composition, while we used a metatranscriptomic approach looking at only expressed genes. In this context it is interesting that other metatranscriptomic results (4) agree with ours in that it was mainly microbes of low occurrence that differed between the preterm and term group.

Overall, the lack of clear differences in the transcriptionally active vaginal microbiome between women with term and preterm pregnancies, suggests that the metatranscriptome has a limited ability to serve as a diagnostic tool for identification of those at high risk for preterm delivery.

Great job Tove and the rest of the team! It was a pleasure working with all of you! The entire paper can be read here.

References

1. Wikström T, Abrahamsson S, Bengtsson-Palme J, Ek CJ, Kuusela P, Rekabdar E, Lindgren P, Wennerholm UB, Jacobsson B, Valentin L, Hagberg H: Microbial and human transcriptome in vaginal fluid at midgestation: association with spontaneous preterm delivery. Clinical and Translational Medicine, 12, 9, e1023 (2022). doi: 10.1002/ctm2.1023 [Paper link]
2. Kindinger LM, Bennett PR, Lee YS, et al.: The interaction between vaginal microbiota, cervical length, and vaginal progesterone treatment for preterm birth risk. Microbiome, 5, 1, 1-14 (2017).
3. Tabatabaei N, Eren AM, Barreiro LB, et al.: Vaginal microbiome in early pregnancy and subsequent risk of spontaneous preterm birth: a case-control study. BJOG, 126, 3, 349-358 (2019).
4. Fettweis JM, Serrano MG, Brooks JP, et al.: The vaginal microbiome and preterm birth. Nature Medicine, 25, 6, 1012-1021 (2019).

Last week, a paper resulting from a collaboration with Stefanie Heß and Viktor Jonsson was published in Environmental Science & Technology. In the paper, we build a quantitative model for the emergence of antibiotic resistance genes in human pathogens and populate it using the few numbers that are available on different processes (bacterial uptake, horizontal gene transfer rates, rate of mobilization of chromosomal genes, etc.) in the literature (1).

In short, we find that in order for the environment to play an important role in the appearance of novel resistance genes in pathogens, there needs to be a substantial flow of bacteria from the environment to the human microbiome. We also find that most likely the majority of resistance genes in human pathogens have very small fitness costs associated with them, if any cost at all.

The model makes three important predictions:

1. The majority of ARGs present in pathogens today should have very limited effects on fitness. The model caps the average fitness impact for ARGs currently present in human pathogens between −0.2 and +0.1% per generation. By determining the fitness effects of carrying individual ARGs in their current hosts, this prediction could be experimentally tested.
2. The most likely location of ARGs 70 years ago would have been in human-associated bacteria. By tracking ARGs currently present in human pathogens across bacterial genomes, it may be possible to trace the evolutionary history of these genes and thereby identify their likely hosts at the beginning of the antibiotic era, similar to what was done by Stefan Ebmeyer and his colleagues (2). What they found sort-of corroborate the findings of our model and lend support to the idea that most ARGs may not originate in the environment. However, this analysis is complicated by the biased sampling of fully sequenced bacterial genomes, most of which originate from human specimens. That said, the rapid increase in sequencing capacity may make full-scale analysis of ARG origins using genomic data possible in the near future, which would enable testing of this prediction of the model.
3. If the origins of ARGs currently circulating in pathogens can be established, the range of reasonable dispersal ability levels from the environment to pathogens narrows dramatically. Similarly, if the rates of mobilization and horizontal transfer of resistance genes could be better determined by experiments, the model would predict the likely origins more precisely. Just establishing a ball-park range for the mobilization rate would dramatically restrict the possible outcomes of the model. Thus, a more precise determination of any of these parameters would enable several more specific predictions by the model.

This paper has a quite interesting backstory, beginning with me having leftover time on a bus ride in Madison (WI), thinking about whether you could quantize the conceptual framework for resistance gene emergence we described in our 2018 review paper in FEMS Reviews Microbiology (3). This spurred the first attempt at such a model, which then led to Stefanie Heß and me applying for support from the Centre for Antibiotic Resistance Research at the University of Gothenburg (CARe) to develop this idea further. We got this support and Stefanie spent a few days with me in Gothenburg developing this idea into a model we could implement in R.

However, at that point we realized we needed more modeling expertise and brought in Viktor Jonsson to make sure the model was robust. From there, it took us about 1.5 years to refine and rerun the model about a million times… By the early spring this year, we had a reasonable model that we could write a manuscript around, and this is what now is published. It’s been an interesting and very nice ride together with Stefanie and Viktor!

References

1. Bengtsson-Palme J, Jonsson V, Heß S: What is the role of the environment in the emergence of novel antibiotic resistance genes? A modelling approach. Environmental Science & Technology, Article ASAP (2021). doi: 10.1021/acs.est.1c02977 [Paper link]
2. Ebmeyer S, Kristiansson E, Larsson DGJ: A framework for identifying the recent origins of mobile antibiotic resistance genes. Communications Biology 4 (2021). doi: 10.1038/s42003-020-01545-5
3. Bengtsson-Palme J, Kristiansson E, Larsson DGJ: Environmental factors influencing the development and spread of antibiotic resistance. FEMS Microbiology Reviews, 42, 1, 68–80 (2018). doi: 10.1093/femsre/fux053 [Paper link]

Last week, I published 13 (!!) papers in the EFSA Journal on how to assess concentrations of antibiotics that could select for antibiotic resistance in animal feed (1-13). Or, well, you could also look at it as that the EFSA Biohaz panel that I have been a part of for more than two years published our final 13-part report.

Regardless of how you view it, this set of papers have two main takeaways:

1. We present a method to establish Predicted Minimal Selective Concentrations (PMSCs) for antibiotics. This method uses a combination of Dan Andersson’s approach to MSCs (14) and the method I published with Joakim Larsson around five years ago to establish predicted no-effect concentrations (PNECs) for antibiotics based on MIC data (15). The combination is a powerful (but very cautious) tool to estimate minimal selective concentrations for antibiotics (1), and we have subsequently applied this method to animal feed contamination with antibiotics, but…
2. There is way too little data to establish PMSCs for most antibiotics with any certainty. Really, the lack of data is so bad that for many of the antibiotic classes we could not make a reasonable assessment. Thus the main conclusion might be that we need a lot more data on how low concentrations of antibiotics that select for antibiotic resistance, both in laboratory systems and in more realistic settings.

References

1. EFSA Panel on Biological Hazards (BIOHAZ)*, Allende A, Koutsoumanis K, Alvarez-Ordóñez A, Bolton D, Bover-Cid S, Chemaly M, Davies R, De Cesare A, Herman L, Hilbert F, Lindqvist R, Nauta M, Ru G, Simmons M, Skandamis P, Suffredini E, Andersson DI, Bampidis V, Bengtsson-Palme J, Bouchard D, Ferran A, Kouba M, López Puente S, López-Alonso M, Saxmose Nielsen S, Pechová A, Petkova M, Girault S, Broglia A, Guerra B, Lorenzo Innocenti M, Liébana E, López-Gálvez G, Manini P, Stella P, Peixe L: Maximum levels of cross-contamination for 24 antimicrobial active substances in non-target feed. Part 1: Methodology, general data gaps and uncertainties. EFSA Journal, 19, 10 (2021). doi: 10.2903/j.efsa.2021.6852 [Paper link]
2. EFSA Panel on Biological Hazards (BIOHAZ)*, Allende A, Koutsoumanis K, Alvarez-Ordóñez A, Bolton D, Bover-Cid S, Chemaly M, Davies R, De Cesare A, Herman L, Hilbert F, Lindqvist R, Nauta M, Ru G, Simmons M, Skandamis P, Suffredini E, Andersson DI, Bampidis V, Bengtsson-Palme J, Bouchard D, Ferran A, Kouba M, López Puente S, López-Alonso M, Saxmose Nielsen S, Pechová A, Petkova M, Girault S, Broglia A, Guerra B, Lorenzo Innocenti M, Liébana E, López-Gálvez G, Manini P, Stella P, Peixe L: Maximum levels of cross-contamination for 24 antimicrobial active substances in non-target feed. Part 2: Aminoglycosides/aminocyclitols: apramycin, paromomycin, neomycin and spectinomycin. EFSA Journal, 19, 10 (2021). doi: 10.2903/j.efsa.2021.6853 [Paper link]
3. EFSA Panel on Biological Hazards (BIOHAZ)*, Allende A, Koutsoumanis K, Alvarez-Ordóñez A, Bolton D, Bover-Cid S, Chemaly M, Davies R, De Cesare A, Herman L, Hilbert F, Lindqvist R, Nauta M, Ru G, Simmons M, Skandamis P, Suffredini E, Andersson DI, Bampidis V, Bengtsson-Palme J, Bouchard D, Ferran A, Kouba M, López Puente S, López-Alonso M, Saxmose Nielsen S, Pechová A, Petkova M, Girault S, Broglia A, Guerra B, Lorenzo Innocenti M, Liébana E, López-Gálvez G, Manini P, Stella P, Peixe L: Maximum levels of cross-contamination for 24 antimicrobial active substances in non-target feed. Part 3: Amprolium. EFSA Journal, 19, 10 (2021). doi: 10.2903/j.efsa.2021.6854 [Paper link]
4. EFSA Panel on Biological Hazards (BIOHAZ)*, Allende A, Koutsoumanis K, Alvarez-Ordóñez A, Bolton D, Bover-Cid S, Chemaly M, Davies R, De Cesare A, Herman L, Hilbert F, Lindqvist R, Nauta M, Ru G, Simmons M, Skandamis P, Suffredini E, Andersson DI, Bampidis V, Bengtsson-Palme J, Bouchard D, Ferran A, Kouba M, López Puente S, López-Alonso M, Saxmose Nielsen S, Pechová A, Petkova M, Girault S, Broglia A, Guerra B, Lorenzo Innocenti M, Liébana E, López-Gálvez G, Manini P, Stella P, Peixe L: Maximum levels of cross-contamination for 24 antimicrobial active substances in non-target feed. Part 4: ß-Lactams: amoxicillin and penicillin. EFSA Journal, 19, 10 (2021). doi: 10.2903/j.efsa.2021.6855 [Paper link]
5. EFSA Panel on Biological Hazards (BIOHAZ)*, Allende A, Koutsoumanis K, Alvarez-Ordóñez A, Bolton D, Bover-Cid S, Chemaly M, Davies R, De Cesare A, Herman L, Hilbert F, Lindqvist R, Nauta M, Ru G, Simmons M, Skandamis P, Suffredini E, Andersson DI, Bampidis V, Bengtsson-Palme J, Bouchard D, Ferran A, Kouba M, López Puente S, López-Alonso M, Saxmose Nielsen S, Pechová A, Petkova M, Girault S, Broglia A, Guerra B, Lorenzo Innocenti M, Liébana E, López-Gálvez G, Manini P, Stella P, Peixe L: Maximum levels of cross-contamination for 24 antimicrobial active substances in non-target feed. Part 5: Lincosamides: lincomycin. EFSA Journal, 19, 10 (2021). doi: 10.2903/j.efsa.2021.6856 [Paper link]
6. EFSA Panel on Biological Hazards (BIOHAZ)*, Allende A, Koutsoumanis K, Alvarez-Ordóñez A, Bolton D, Bover-Cid S, Chemaly M, Davies R, De Cesare A, Herman L, Hilbert F, Lindqvist R, Nauta M, Ru G, Simmons M, Skandamis P, Suffredini E, Andersson DI, Bampidis V, Bengtsson-Palme J, Bouchard D, Ferran A, Kouba M, López Puente S, López-Alonso M, Saxmose Nielsen S, Pechová A, Petkova M, Girault S, Broglia A, Guerra B, Lorenzo Innocenti M, Liébana E, López-Gálvez G, Manini P, Stella P, Peixe L: Maximum levels of cross-contamination for 24 antimicrobial active substances in non-target feed. Part 6: Macrolides: tilmicosin, tylosin and tylvalosin. EFSA Journal, 19, 10 (2021). doi: 10.2903/j.efsa.2021.6858 [Paper link]
7. EFSA Panel on Biological Hazards (BIOHAZ)*, Allende A, Koutsoumanis K, Alvarez-Ordóñez A, Bolton D, Bover-Cid S, Chemaly M, Davies R, De Cesare A, Herman L, Hilbert F, Lindqvist R, Nauta M, Ru G, Simmons M, Skandamis P, Suffredini E, Andersson DI, Bampidis V, Bengtsson-Palme J, Bouchard D, Ferran A, Kouba M, López Puente S, López-Alonso M, Saxmose Nielsen S, Pechová A, Petkova M, Girault S, Broglia A, Guerra B, Lorenzo Innocenti M, Liébana E, López-Gálvez G, Manini P, Stella P, Peixe L: Maximum levels of cross-contamination for 24 antimicrobial active substances in non-target feed. Part 7: Amphenicols: florfenicol and thiamphenicol. EFSA Journal, 19, 10 (2021). doi: 10.2903/j.efsa.2021.6859 [Paper link]
8. EFSA Panel on Biological Hazards (BIOHAZ)*, Allende A, Koutsoumanis K, Alvarez-Ordóñez A, Bolton D, Bover-Cid S, Chemaly M, Davies R, De Cesare A, Herman L, Hilbert F, Lindqvist R, Nauta M, Ru G, Simmons M, Skandamis P, Suffredini E, Andersson DI, Bampidis V, Bengtsson-Palme J, Bouchard D, Ferran A, Kouba M, López Puente S, López-Alonso M, Saxmose Nielsen S, Pechová A, Petkova M, Girault S, Broglia A, Guerra B, Lorenzo Innocenti M, Liébana E, López-Gálvez G, Manini P, Stella P, Peixe L: Maximum levels of cross-contamination for 24 antimicrobial active substances in non-target feed. Part 8: Pleuromutilins: tiamulin and valnemulin. EFSA Journal, 19, 10 (2021). doi: 10.2903/j.efsa.2021.6860 [Paper link]
9. EFSA Panel on Biological Hazards (BIOHAZ)*, Allende A, Koutsoumanis K, Alvarez-Ordóñez A, Bolton D, Bover-Cid S, Chemaly M, Davies R, De Cesare A, Herman L, Hilbert F, Lindqvist R, Nauta M, Ru G, Simmons M, Skandamis P, Suffredini E, Andersson DI, Bampidis V, Bengtsson-Palme J, Bouchard D, Ferran A, Kouba M, López Puente S, López-Alonso M, Saxmose Nielsen S, Pechová A, Petkova M, Girault S, Broglia A, Guerra B, Lorenzo Innocenti M, Liébana E, López-Gálvez G, Manini P, Stella P, Peixe L: Maximum levels of cross-contamination for 24 antimicrobial active substances in non-target feed. Part 9: Polymyxins: colistin. EFSA Journal, 19, 10 (2021). doi: 10.2903/j.efsa.2021.6861 [Paper link]
10. EFSA Panel on Biological Hazards (BIOHAZ)*, Allende A, Koutsoumanis K, Alvarez-Ordóñez A, Bolton D, Bover-Cid S, Chemaly M, Davies R, De Cesare A, Herman L, Hilbert F, Lindqvist R, Nauta M, Ru G, Simmons M, Skandamis P, Suffredini E, Andersson DI, Bampidis V, Bengtsson-Palme J, Bouchard D, Ferran A, Kouba M, López Puente S, López-Alonso M, Saxmose Nielsen S, Pechová A, Petkova M, Girault S, Broglia A, Guerra B, Lorenzo Innocenti M, Liébana E, López-Gálvez G, Manini P, Stella P, Peixe L: Maximum levels of cross-contamination for 24 antimicrobial active substances in non-target feed. Part 10: Quinolones: flumequine and oxolinic acid. EFSA Journal, 19, 10 (2021). doi: 10.2903/j.efsa.2021.6862 [Paper link]
11. EFSA Panel on Biological Hazards (BIOHAZ)*, Allende A, Koutsoumanis K, Alvarez-Ordóñez A, Bolton D, Bover-Cid S, Chemaly M, Davies R, De Cesare A, Herman L, Hilbert F, Lindqvist R, Nauta M, Ru G, Simmons M, Skandamis P, Suffredini E, Andersson DI, Bampidis V, Bengtsson-Palme J, Bouchard D, Ferran A, Kouba M, López Puente S, López-Alonso M, Saxmose Nielsen S, Pechová A, Petkova M, Girault S, Broglia A, Guerra B, Lorenzo Innocenti M, Liébana E, López-Gálvez G, Manini P, Stella P, Peixe L: Maximum levels of cross-contamination for 24 antimicrobial active substances in non-target feed. Part 11: Sulfonamides. EFSA Journal, 19, 10 (2021). doi: 10.2903/j.efsa.2021.6863 [Paper link]
12. EFSA Panel on Biological Hazards (BIOHAZ)*, Allende A, Koutsoumanis K, Alvarez-Ordóñez A, Bolton D, Bover-Cid S, Chemaly M, Davies R, De Cesare A, Herman L, Hilbert F, Lindqvist R, Nauta M, Ru G, Simmons M, Skandamis P, Suffredini E, Andersson DI, Bampidis V, Bengtsson-Palme J, Bouchard D, Ferran A, Kouba M, López Puente S, López-Alonso M, Saxmose Nielsen S, Pechová A, Petkova M, Girault S, Broglia A, Guerra B, Lorenzo Innocenti M, Liébana E, López-Gálvez G, Manini P, Stella P, Peixe L: Maximum levels of cross-contamination for 24 antimicrobial active substances in non-target feed. Part 12: Tetracycline, chlortetracycline, oxytetracycline, and doxycycline. EFSA Journal, 19, 10 (2021). doi: 10.2903/j.efsa.2021.6864[Paper link]
13. EFSA Panel on Biological Hazards (BIOHAZ)*, Allende A, Koutsoumanis K, Alvarez-Ordóñez A, Bolton D, Bover-Cid S, Chemaly M, Davies R, De Cesare A, Herman L, Hilbert F, Lindqvist R, Nauta M, Ru G, Simmons M, Skandamis P, Suffredini E, Andersson DI, Bampidis V, Bengtsson-Palme J, Bouchard D, Ferran A, Kouba M, López Puente S, López-Alonso M, Saxmose Nielsen S, Pechová A, Petkova M, Girault S, Broglia A, Guerra B, Lorenzo Innocenti M, Liébana E, López-Gálvez G, Manini P, Stella P, Peixe L: Maximum levels of cross-contamination for 24 antimicrobial active substances in non-target feed. Part 13: Trimethoprim. EFSA Journal, 19, 10 (2021). doi: 10.2903/j.efsa.2021.6865 [Paper link]
14. Gullberg E, Cao S, Berg OG, Ilbäck C, Sandegren L, Hughes D, et al.: Selection of resistant bacteria at very low antibiotic concentrations. PLoS Pathogens 7, e1002158 (2011). doi: 10.1371/journal.ppat.1002158
15. Bengtsson-Palme J, Larsson DGJ: Concentrations of antibiotics predicted to select for resistant bacteria: Proposed limits for environmental regulation. Environment International, 86, 140-149 (2016). doi: 10.1016/j.envint.2015.10.015 [Paper link]

I am very happy to announce that Emil Burman‘s (doctoral student in the lab) first first-author paper was published today in Frontiers in Microbiology. In this paper (1), we explored how temperature affected the interactions in the model microbial community THOR (2). Somewhat surprisingly, we found that even a small difference in temperature changed the community intrinsic properties (3) of this model community a lot. We furthermore find that changes in growth rates of the members of the community partially explains the changed interaction patterns, but only to some extent. Finally, we also found that biofilm production overall was much higher at lower temperatures (9-15°C) than at room temperature, and that at around 25°C and above the community formed virtually no biofilm.

The temperature range we tested is not unlikely to be encountered when incubating the community in a thermally unregulated environment. Thus, our results show that a high degree of temperature control is crucial between experiments, particularly when reproducing results across different laboratories, equipment, and personnel. This highlights the need for standards and transparency in research on microbial model communities (4).

Another important, related, aspect is that disruptive factors that discriminate against single members of the community are not unique to THOR. Instead, this is likely to be the case for other microbial model (as well as natural communities). Since only a few of these model communities have been elucidated for community behaviors outside of specific culturing conditions they were first contrived under, this may severely limit our view of interactions between microbes to specific laboratory settings. This casts some doubt on the validity of extrapolation from results obtained from microbial model communities. It seems to be important moving forward to establish that community-intrinsic behaviors in model communities are stable in the face of variable environmental conditions, such as temperature, pH, nutrient availability, and initial inoculum size.

A short backstory to this paper: this begun when Emil could not consistently replicate the results I had obtained during my postdoc (working on THOR) in Prof. Jo Handelsman’s lab at the University of Wisconsin-Madison. After a long time of troubleshooting, we realized that our lab did not hold a stable room temperature. We bought a cold incubator, and – boom – after that the expected community behavior came back. This made us realize the importance of temperature for the community-intrinsic properties of THOR, which then led to this more systematic investigation.

Great work Emil! It is nice to finally see this in its published form. Read the entire paper (open access) here!

References

1. Burman E, Bengtsson-Palme J: Microbial community interactions are sensitive to small differences in temperature. Frontiers in Microbiology, 12, 672910 (2021). doi: 10.3389/fmicb.2021.672910
2. Lozano GL, Bravo JI, Garavito Diago MF, Park HB, Hurley A, Peterson SB, Stabb EV, Crawford JM, Broderick NA, Handelsman J: Introducing THOR, a Model Microbiome for Genetic Dissection of Community Behavior. mBio, 10, 2, e02846-18 (2019). doi: 10.1128/mBio.02846-18
3. Madsen JS, Sørensen SJ, Burmølle M: Bacterial social interactions and the emergence of community-intrinsic properties. Current Opinion in Microbiology 42, 104–109 (2018). doi: 10.1016/j.mib.2017.11.018
4. Bengtsson-Palme J: Microbial model communities: To understand complexity, harness the power of simplicity. Computational and Structural Biotechnology Journal, 18, 3987-4001 (2020). doi: 10.1016/j.csbj.2020.11.043

We start the new year with a bang, or at least a new paper published. Bioinformatics put our paper (1) describing the software package CAFE online today (although it was accepted late last year). The CAFE package is a combination of Perl and R tools that can analyze data from paired transposon mutant sequencing experiments (2-4), generate fitness coefficients for each gene and condition, and perform appropriate statistical testing on these fitness coefficients. The paper is short, but shows that CAFE performs as good as the best competing tools (5-7) while being superior at controlling for false positives (you’ll have to dig into the supplement to find the data for that though).

Importantly, this is a collaborative effort by basically the entire research group from last spring: me, Haveela, Emil, Anna and our visiting student Adriana. A big thanks to all of you for working on this short but important paper! You can read the full paper here.

References

1. Abramova A, Osińska A, Kunche H, Burman E, Bengtsson-Palme J (2021) CAFE: A software suite for analysis of paired-sample transposon insertion sequencing data. Bioinformatics, advance article doi: 10.1093/bioinformatics/btaa1086
2. Chao,M.C. et al. (2016) The design and analysis of transposon insertion sequencing experiments. Nature reviews Microbiology, 14, 119–128.
3. van Opijnen,T. and Camilli,A. (2013) Transposon insertion sequencing: a new tool for systems-level analysis of microorganisms. Nature reviews Microbiology, 11, 435–442.
4. Goodman,A.L. et al. (2011) Identifying microbial fitness determinants by insertion sequencing using genome-wide transposon mutant libraries. Nature Protocols, 6, 1969–1980.
5. McCoy,K.M. et al. (2017) MAGenTA: a Galaxy implemented tool for complete Tn- Seq analysis and data visualization. Bioinformatics, 33, 2781– 2783.
6. Zhao,L. et al. (2017) TnseqDiff: identification of conditionally essential genes in transposon sequencing studies. BMC Bioinformatics, 18.
7. Zomer,A. et al. (2012) ESSENTIALS: Software for Rapid Analysis of High Throughput Transposon Insertion Sequencing Data. PLoS ONE, 7, e43012.

This week, in a stroke of luck coinciding with my conference presentation on the same topic, my review paper on microbial model communities came out in Computational and Structural Biotechnology Journal. The paper (1) provides an overview of the existing microbial model communities that have been developed for different purposes and makes some recommendations on when to use what kind of community. I also make a deep-dive into community intrinsic-properties and how to capture and understand how microbes growing together interact in a way that is not predictable from how they grow in isolation.

The main take-home messages of the paper are that 1) there already exists a quite diverse range of microbial model communities – we probably don’t need a wealth of additional model systems, 2) there need to be better standardization and description of the exact protocols used – this is more important in multi-species communities than when species are grown in isolation, and 3) the researchers working with microbial model communities need to settle on a ‘gold standard’ set of model communities, as well as common definitions, terms and frameworks, or the complexity of the universe of model systems itself may throw a wrench into the research made using these model systems.

The paper was inspired by the work I did in Jo Handelsman‘s lab on the THOR model community (2), which I then have brought with me to the University of Gothenburg. In the lab, we are also setting up other model systems for microbial interactions, and in this process I thought it would be useful to make an overview of what is already out there. And that overview then became this review paper.

The paper is fully open-access, so there is really not much need to go into the details here. Go and read the entire thing instead (or just get baffled by Table 1, listing the communities that are already out there!)

References

1. Bengtsson-Palme J: Microbial model communities: To understand complexity, harness the power of simplicity. Computational and Structural Biotechnology Journal, in press (2020). doi: 10.1016/j.csbj.2020.11.043
2. Lozano GL, Bravo JI, Garavito Diago MF, Park HB, Hurley A, Peterson SB, Stabb EV, Crawford JM, Broderick NA, Handelsman J: Introducing THOR, a Model Microbiome for Genetic Dissection of Community Behavior. mBio, 10, 2, e02846-18 (2019). doi: 10.1128/mBio.02846-18