Discovery of antimicrobial peptides in the global microbiome with machine learning.

Novel antibiotics are urgently needed to combat the antibiotic-resistance crisis. We present a machine-learning-based approach to predict antimicrobial peptides (AMPs) within the global microbiome and leverage a vast dataset of 63,410 metagenomes and 87,920 prokaryotic genomes from environmental and host-associated habitats to create the AMPSphere, a comprehensive catalog comprising 863,498 non-redundant peptides, few of which match existing databases. AMPSphere provides insights into the evolutionary origins of peptides, including by duplication or gene truncation of longer sequences, and we observed that AMP production varies by habitat. To validate our predictions, we synthesized and tested 100 AMPs against clinically relevant drug-resistant pathogens and human gut commensals both in vitro and in vivo. A total of 79 peptides were active, with 63 targeting pathogens. These active AMPs exhibited antibacterial activity by disrupting bacterial membranes. In conclusion, our approach identified nearly one million prokaryotic AMP sequences, an open-access resource for antibiotic discovery.

Computational exploration of the global microbiome for antibiotic discovery.

Novel antibiotics are urgently needed to combat the antibiotic-resistance crisis. We present a machine learning-based approach to predict prokaryotic antimicrobial peptides (AMPs) by leveraging a vast dataset of 63,410 metagenomes and 87,920 microbial genomes. This led to the creation of AMPSphere, a comprehensive catalog comprising 863,498 non-redundant peptides, the majority of which were previously unknown. We observed that AMP production varies by habitat, with animal-associated samples displaying the highest proportion of AMPs compared to other habitats. Furthermore, within different human-associated microbiota, strain-level differences were evident. To validate our predictions, we synthesized and experimentally tested 50 AMPs, demonstrating their efficacy against clinically relevant drug-resistant pathogens both in vitro and in vivo. These AMPs exhibited antibacterial activity by targeting the bacterial membrane. Additionally, AMPSphere provides valuable insights into the evolutionary origins of peptides. In conclusion, our approach identified AMP sequences within prokaryotic microbiomes, opening up new avenues for the discovery of antibiotics.

Relationship between 233 colorectal cancer risk loci and survival in 1926 patients with advanced disease.

Background: Genome, transcriptome and methylome-wide association studies have identified single-nucleotide polymorphisms (SNPs) or genes at 258 loci associated with colorectal cancer (CRC) risk. We studied the relationship between these and patient outcome. Methods: We studied 1926 unrelated patients with advanced CRC from COIN and COIN-B. Of 205 CRC-risk SNPs, 19 were directly genotyped and 162 were imputed, and of 53 risk genes, 52 were tested. An additive Cox model for overall survival was adjusted for known prognostic factors. For nominally significant SNPs or genes, we considered a recessive model with a Bonferroni corrected threshold of P = 2.1 × 10-4. We examined SNPs as expression quantitative trait loci (eQTL) and the relationship between gene expression in colorectal tumours and survival in 597 unrelated patients. Results: Eleven SNPs or genes were nominally associated with survival under an additive model. Only rs117079142 mapping to UTP23 and EIF3H (Hazard Ratio [HR] = 2.79, 95% Confidence Intervals [CI] = 1.70-4.58, P = 4.7 × 10-5) and rs9924886 mapping to CDH1 and CDH3 (HR = 1.24, 95% CI = 1.12-1.38, P = 5.2 × 10-5) passed the multiple testing threshold under a recessive model. rs117079142 was an eQTL for UTP23 and rs9924886 for CDH1, CDH3 and ZFP90. Decreased CDH1 expression in CRCs was associated with worse survival (HR = 2.18, 95% CI = 1.3-3.5, P = 1.8 × 10-3). Conclusion: rs117079142 and rs9924886 may represent potential prognostic biomarkers for CRC.