Genomics seminars

Laboratory-evolved gentamicin-resistant bacteria are analysed using whole-genome sequencing and protein-structure modelling. The study focuses on mutations in elongation factor G and compares AlphaFold2 predictions with wild-type proteins. Aligning predicted EF-G structures with experimental ribosome and antibiotic-bound structures reveals how local conformational and physicochemical changes could affect drug binding. The talk examines subtle structural shifts, mutations without detectable structural changes, and the limitations of prediction alone. It presents a workflow that combines genomic evidence, AI models and experimental structures to form mechanistic hypotheses about antimicrobial resistance.

Genome annotation connects assembled sequences with functional information, but many sequenced genomes remain unannotated and large-scale annotation is difficult to standardise and evaluate. The Australian Tree of Life Bioinformatics Group and international biodiversity researchers are working to automate genome assembly and annotation so that researchers without specialist bioinformatics expertise can use them. During an industry placement hosted by the Australian Genome Research Facility with Australian BioCommons, University of Queensland PhD candidate Jane Tung compared current annotation pipelines across a range of genomes to identify scalable approaches for Australian reference genomes. This webinar presents the benchmarking findings and examines how they can inform and streamline genome assembly and annotation workflows. It is intended for researchers using genomes and genome annotations.

MGnify supports microbiome research through versioned pipelines for taxonomic and functional analysis and access to previously analysed public and submitted datasets. This introduction explains the resource’s data types—amplicons, raw reads, assemblies, genomes and proteins—and how studies, samples and sequencing runs connect to analyses, visualisations and downloadable results. It covers browsing and programmatic access for further exploration and introduces the subsequent metagenomics webinar series. The session is intended for life-science students and researchers; familiarity with DNA sequencing helps, while specialist coding or bioinformatics experience is unnecessary.

Three presentations examine how long-read sequencing can improve biodiversity resources and their use in conservation and environmental decisions. Evgeny Glazov introduces amplification-free nanopore sequencing, direct detection of base modifications and flexible read lengths for natural-resource research. Kara Dicks discusses scaling bird and fish reference assemblies through the Tree of Life programme. The presentation addresses completeness of genes, repeats and regulatory regions, difficult avian microchromosomes and GC-rich sequences, and the balance between standardized production and adaptation to diverse species. Steven Robbins examines marine microbes missed by short-read assembly because of GC bias and strain diversity. Work underlying the Great Barrier Reef Microbial Genomes Database uses long-read and hybrid assemblies to recover overlooked lineages including Pelagibacter, Prochlorococcus and SAR86. The reported improvements include twice the genome recovery, three times as many high-quality assemblies and approximately tenfold greater contiguity. Long reads also expose chromosome-scale algal genomes and previously missed structural variation. Together, these approaches support biodiversity monitoring, ecosystem assessment and policy.

Protein-structure prediction can help distinguish competing gene models. Studies of two fungi and an apicoplast compare original and manually curated annotations, finding that structure-derived scores favour curated models when predictions differ in structured domains. These signals complement sequence-domain analysis but offer little help for uncertainty in potentially disordered regions. The presentation also examines Protenix-mini as a faster alternative retaining discriminatory power, complete-proteome scoring, and a NextFlow pipeline intended to make the workflow usable. It considers applications when transcriptomic evidence is limited and the remaining limitations of structure-based annotation.

Josh Levendis examines how N6-methyladenosine RNA modification controls gene expression in Plasmodium falciparum. Nanopore direct RNA sequencing and GLORI-seq reveal enrichment near transcript ends. Inducible mislocalisation of methyltransferase depletes this modification and produces readthrough and chimeric transcripts, implicating methylation in correct polyadenylation and transcription termination. Stage-dependent modification of exons and retained introns may also regulate early termination during sexual development. Alex Harris investigates why protection from RTS,S malaria vaccination and seasonal chemoprevention declines despite annual boosters. Analyses of plasma from 1,929 West African children followed over four years assess antibody responses to different antigen regions and their variation with host factors, coinfections and micronutrients. The work examines declining booster responses and aims to inform strategies for durable vaccine protection.

Interpreting human missense variants requires connecting genomic coordinates with protein sequences, structures and functional evidence. ProtVar integrates annotations and predictions from UniProt, Ensembl, PDBe, Open Targets and AlphaFoldDB, extending variant lookup to discovery across more than 500 million potential missense variants. This seminar demonstrates natural-language searches using diseases, pathways and drug responses, multi-criteria filtering, structural visualisation and data export. Participants will learn to prioritise variants, interpret their structural and functional effects, and retrieve data through web downloads or the REST API. The session serves geneticists, drug-discovery researchers and computational biologists; undergraduate molecular biology and genetics are recommended.

Bacteriophage genomes offer a large reservoir of enzymes adapted to extreme environments. This seminar examines endolysins, enzymes that break down bacterial cell walls, as potential antimicrobial agents for agriculture, food production and sanitation. Cold-adapted variants could support cleaning at lower temperatures, while heat-tolerant variants may serve fermentation, food processing and alternatives to antibiotics. Amy Wiseman presents research combining metagenomic sequences from Icelandic hot springs and Arctic seawater with AlphaFold and other prediction methods, biophysical experiments, X-ray crystallography and biochemical characterization. The work investigates candidate cold-adapted and thermophilic endolysins, reveals unexpected structural features, and considers what these features imply for evolution, environmental adaptation and enzyme function. The talk connects sequence mining and structural prediction with experimental validation, biotechnology applications and rational enzyme design.