**Standard disclaimer: these are my notes taken during the sessions, as accurate as I can make them, please let me know if any inaccuracies. I’m putting them here as others have said they found them useful. However worth re-checking anything before you re-quote it. If this continues to be useful, I’ll continue. If it’s clear i’m not reaching the standards of accuracy I’m aiming for, I’ll review future posting plans.**
Final day at ECCMID! There was a very high quality Microbiome session, with two outstanding talks from the speakers, as overviews on the applications of microbiome research – highly recommended viewing and links below.
The massive highlight of the genomics session for me was a presentation by one of our collaborators – Phelim Bradley presenting the work done in the Iqbal group and Mykrobe – a tool that uses WGS data to predict antibiotic resistance. I’ll be right up and admit I work with these chaps, and I don’t want to appear biased, but I can only say that the first time I saw this working a few weeks ago I went Oh My GOD…
As fundamentally a clinician taking a ride in the fun world of genomics, this is the first time I’ve seen something involving genomic analysis I can see myself using on the ward. The user interface looked straight out of Apple & something I could teach my granny (/my senior much-respected Consultant population) how to use. And validated on over 1000 Staph. aureus samples, with major and very major error rates <1% compared with culture phenotype. TB and Gram-neg versions also developed. For any clinician agnostic about whether WGS is ready for direct use in patient treatment, the link for the 10 min presentation available at ECCMID live : here ( -worth a look even if you just skip to the demo!)
Updated: for a view of the granny-friendly Apple-esque look – slides up at slideshare:
(ok maybe not granny, unless she’s a Microbiologically-literate elder, but given I’ve spent the last year in a field where people tell you they’ve written a really user-friendly genomic analysis interface which is super easy to use with just basic python scripting – I’m very grateful to any bioinformaticians who consider the intended end-user of their product… )
A very good question from the audience – given cost of sequencing and current times taken to extract DNA/analyse – how can it be used to treat patients (and what benefits over current clinical workflow). Agree – for that reason, I think at the present moment in time for routine diagnostics, in Staph. aureus it certainly would be hard to justify. But given TB takes weeks/months to phenotype, and the cost of sequencing is going down (and nanopore is looking increasingly promising…) incredibly exciting times.
Previous sessions posited that the use of WGS in day-to-day clinical practice was probably a few years away, and also suggested a key challenge is presenting WGS analysis in a way that can be used by non-specialists. The Mykrobe presentation suggested that perhaps we’re really much closer than previously thought…
Rest of the summaries below.
In terms of the rest of the genomics session – was a fairly broad mix, from groups doing advanced genomic pathogen analysis to smaller projects just starting out.
Johansen – presented work on how genomics can provide insights on how Pseudomonas aeruginosa colonises the lungs of cystic fibrosis patients. (Marvig Sommer Nat Genetics 2015) Extensive work and impressive patient cohort -BALs taken from children who had never had Pseudomonas, and then over multiple time periods, with a total of 474 genomes sequenced. They saw many different patterns of acquision, from either one of the dominant clones, or a previously unseen clone ?from the environment, and were 4 patterns:
1) Persistent culture of one clone in patient following colonisation, often associated with very little immune response, and deterioration of lung function
2) Persistent clone, but occasional ‘invader’
3) Switchers- succession of clone types
4) Multiple clone types, varying numbers, acquisitions losses.
Some thought provoking discussion on their sampling methods, given the increasing understanding of the ‘cloud’ of within-host diversity. The speaker said they sampled multiple colonies if they noted different antibiograms or differing growth (?morphology?) but highlights the need for all groups to consider this if they wish to use WGS to rule in/rule out transmission.
edit: think I might have slightly misunderstood this – went back and watched presentation on ECCMID live ! Better summary now included. (not about mutation rates, about diversity).
Addressing a really important question – studying the diversity in a large database of TB genomes. Identified genes inTB Dream and BIGS DB Database, and separated the genes up into function, (metabolic virulence etc.) and then plotting diversity – showing that some genes very highly conserved, and likely under stabilising selection, and some highly diverse. Basically, one gene (and location on the gene) is not like another! . And highlights limitations of algorithms that make the assumption that one SNP is just like another….
A group from Spain (presenter – Rodriguez Fernandez) showed a really novel approach to defining and visualising the core and accessory genome of E.coli ST131. They took a representative sample of bloodstream sample, randomly selected and used their PLACNET pipeline (Lanza PlosGEn2015) and a newly developed AccNET (Accessory Constellation Network) to first group orthologous clusters and define the core genome of their dataset, and then the accessory genes, visualising these in 2d space next to a phylogenetic tree. This looked VERY COOL, clever stuff, and in the era of resistance transmission by MGE, very much needed.
If you’re interested, these were just really interesting, well delivered talks, so for anyone interested I’d say go watch them now when you have a spare half hour and a cup of tea.
No doubt there will be those who have different opinions, or disagree with some of the conclusions, but as an engaging overview I thought they were great.
Roghmann started with the excellent CDC infographic on how resistance spreads (though it shows a model more in keeping with Enterobacteriaceae than for instance TB). After years of trying to explain this I was really glad to find this a few months ago.
A key hilight that we should always be aware of the collateral damage to the microbiome when we use antibiotics, and the SATURN study, set up to study the effect of antibiotics on carriage of resistant isolates, and one of the most wonderfully convoluted acronyms I’ve seen (Specific Antibiotic Therapies on the prevalence of hUman host resistaNt bacteria…)
Good to hear that chicken probiotics seem to be moving forward – an example of a probiotic given in feed, and has been shown to clear Campylobacter of chicken jejunums. As it was also shown to promote chick growth and health it seems it is increasingly being used in the food industry, and the hope is that it reduces the need to use antibiotics to clear flock colonisation. Finally some discussion on faecal transplants, and a growing body of evidence that faecal transplants can be used to clear chronic ESBL/CRE/VRE colonisation in patients who have recurrent serious infection.
Next was Stephan Harbarth on the times where microbiome manipulation goes wrong. Specifically decontamination – including ICU-based selective gut decontamination. He noted that the Price Meta-analysis of SDD found that use of chlorhexidine washes were associated with increased mortality, with speculation that this may have been due to unintended microbiome consequences, possibly reducing colonisation resistance to multi-drug-resistant Enterobacteriaceae, as previously shown by Popovich et al. He also noted that there’s still a lot we need to understand about faecal transplants, and whilst they show huge promise, we must be vigilant for any ‘unintended consequences’, given the interplay with metabolism and neuro-endocrine systems which have already been demonstrated (My Microbiome Made me do it….)
All in all, great talks.