BDB-Lab Autumn 2023 Updates
Small proteins on the move...
Looking forward.
15-20 Oct. Luis will attend the International Peptide Symposium 2023 in Brisbane and present the talk “A survey of antimicrobial peptides from the global microbiome”.
12-15 Nov. Luis will attend incob23 in Brisbane and present a talk on SemiBin and SemiBin2.
4-8 Dec. ABACBS2023 in Brisbane. The programme is still not out, but Luis will attend.
Feel free to contact Luis if you will be present at any of these venues.
23 Oct. Online SemiBin tutorial. This will be during a US+Australia friendly time (check your timezone). Registration is free, but required.
6 Nov. Online Jug tutorial. As above, is a US+Australia friendly time (check your timezone). Registration is free, but required.
Focus of the Quarter:
Computational exploration of the global microbiome for antibiotic discovery
Célio Dias Santos-Júnior*, Marcelo Der Torossian Torres*, Yiqian Duan, Álvaro Rodríguez del Río, Thomas S.B. Schmidt, Hui Chong, Anthony Fullam, Kuhn Michael, Chengkai Zhu, Amy Houseman, Jelena Somborski, Anna Vines, Xing-Ming Zhao, Peer Bork, Jaime Huerta-Cepas, Cesar de la Fuente-Nunez⁺, Luis Pedro Coelho⁺ in bioRxiv 2023.08.31.555663; doi: https://doi.org/10.1101/2023.08.31.555663
The fast growth of the number of infections by resistant microorganisms makes the discovery of novel antibiotics urgently needed, and antimicrobial peptides (AMPs) are one source of novel molecules. In addition to clinical use, AMPs can be used in the food industry, veterinary sector, agriculture, or others. Most often, they target microbial membranes, and therefore, AMPs can have different levels of selectivity and sensitivity. Although interesting, few AMPs are known, and most of the known ones are from Eukaryotes. We had previously developed Macrel (Santos-Júnior et al., 2020), an algorithm that predicts AMPs from (meta)genomic samples. Macrel was trained with a realistic proportion of positive and negative examples, and thus, its model returns fewer false positives than competing tools.
Using this machine learning approach, we leveraged a vast dataset of 63,410 metagenomes and 87,920 microbial genomes to create the AMPSphere, a comprehensive catalog comprising 863,498 non-redundant peptides, the majority of which were previously unknown. Now, the description of this work is available as a preprint in BioRXiv. This work was the result of a collaboration with the fantastic de la Fuente group at UPenn and also includes contributions from groups in Madrid and the EMBL.
By using AMPSphere, we detected some interesting ecological effects of AMP production across the different habitats available in our dataset, such as its habitat specificity. We also gained insights into the evolution of the AMPs, observing gene truncation and gene duplication as viable mechanisms for their generation.
Finally, we validated some of our predictions by synthesizing and experimentally testing 50 AMPs, demonstrating the efficacy of these candidates against clinically relevant drug-resistant pathogens both in vitro and in vivo. Some of the tested candidates were also effective against commensals of human guts, which is interesting for microbiome engineering approaches in the human microbiota. In a general way, the tested AMPs exhibited antibacterial activity by targeting the bacterial membrane, a similar mechanism to those previously described. Altogether, AMPSphere will open up new avenues for discovering antibiotics with different ranges of targets and possibly new applications.
People.
Marija is now starting a postdoc at the Sunagawa lab at ETH Zurich. She will work on several projects within the NCCR Microbiomes consortium and take over feature development for the new version of mOTUs.
Congratulations to Vedanth, who got a grant from Emergent Ventures India, and will visit us in Brisbane in 2024.
Congratulations to Luis on earning an ARC Future Fellowship. The project “Understanding prokaryotic small proteins from context” aims to develop new methods to identify and predict the functions of small proteins from microbial communities by studying sequence patterns in their genomes. Luis officially started working at the Center of Microbiome Research at Queensland University of Technology this September.
Apply for a Ph.D. with Luis in Centre for Microbiome Research (Brisbane, Australia). There are funded studentships tied to the Future Fellowship, focusing on identifying and characterizing the small proteins of the global microbiome, but also a second project that will focus on developing novel bioinformatic methods for studying microbial communities.
Manuscripts.
SemiBin2: self-supervised contrastive learning leads to better MAGs for short- and long-read sequencing is finally out in Bioinformatics.
Towards monitoring of antimicrobial resistance in the environment: For what reasons, how to implement it, and what are the data needs? was recently published in Environment International.
A global survey of eco-evolutionary pressures acting on horizontal gene transfer is now available as a preprint in Research Gate.
Computational exploration of the global microbiome for antibiotic discovery is now available as a preprint in Biorxiv (see above).
Read Svetlana’s advice on making the most of remote events: How remote conferencing broadened my horizons and opened career paths (Nature Careers column).
For long-term sustainable software in bioinformatics originated as follow-up discussions to our Summer 2022 Update and has been submitted for publication as a perspective piece (no preprint as it is not original research).
Conferences/talks.
Anna attended the FEMS2023 in Hamburg (Germany) and presented her work on dog gut microbiome.
Luis attended the ISMB/ECCB in Lyon (France) and presented the SemiBin2 paper (YouTube re-recording and slides [press ‘?’ to see key shortcuts]).
Luis visited Norwich Research Park and gave a talk at the Earlham Institute.
Our YouTube channel has several new videos, which we won’t exhaustively list here.