Metagenomics of the Human Intestinal Tracthttp://www.metahit.eu

ISAPP 2010, Barcelona, SpainS. Dusko Ehrlich

Metagenome can be defined as the ensemble of genes & genomes of the

microbes from a given ecological niche.Metagenomics is a powerful approach to characterize composition, properties

and dynamics of the microbiome via metagenome.

The human intestinal microbiota is a forgotten organ…

100 trillion microorganisms ; 10-fold more cells than the human body; 2 kg of mass!Interface between food and epithelium In contact with the 1st pool of immune cells and the 2nd pool of neural cells of the body

…with a major role in health & disease !

Frailty in seniors Van Tongeren et al., 2005Crohn Seksik et al., 2003; Sokol et al., 2006, 2008, 2009Ulcerative colitis Sokol et al., 2008; Martinez et al., 2008Pouchitis Lim et al., 2009, Kühbacher et al., 2006Obesity Ley et al., 2007; Kalliomäki et al., 2008Type-2 diabetes Cani and Delzenne, 2009Type-1 diabetes Dessein et al., 2009; Wen et al., 2008Coeliac disease Nadal et al., 2007; Collado et al., 2009Allergy Kirjavainen et al., 2002; Björkstén, 2009Autism Finegold et al., 2002; Paracho et al., 2005Colorectal cancer Mai et al., 2007; Scanlan et al., 2008 HIV infection Gori et al., 2008

Perturbation of intestinal microbiota as a possible disease factor

* Indications from animal models, effects of antibiotics or probiotics, clinical studies; courtesy of Joël Doré

EU: MetaHIT, 20 M €Ireland: ElderMET, 4 M €Japan: Human Metagenome Consortium, 5 M $Sino-French cooperation: Micro-Obese : 4 M €USA: NIH Human Microbiome Project, 115 M $

The Crohn's & Colitis Foundation of America Gut Microbiome Initiative: Washington U., St. Louis & U. of Denver, Boulder

Worldwide movement to characterise human microbiome

Canada: Microbiome initiative, 14 M $ (August 10, 2010)MSU Wins $7.3M Gut Microbiome Grant | GenomeWeb Daily News | GenomeWeb (August 20, 2010)

International Human Microbiome Consortium

Full MembersCanadian Institutes of Health Research (Canada)European Commission (Europe)Institut National de la Recherche Agronomique (France) Japan Science & Technology Agency, JST (Japan)National Institutes of Health (United States of America)Commonwealth Scientific and Industrial Research Organisation (Australia)Medical Research Council (Gambia)Observing MembersGenome Canada (Canada)Ministry of Education, Cultures, Sports, Sciences and Technology, MEXT (Japan)National Health and Medical Research Council (Australia)Korea Research Institute of Bioscience and Biotechnology (Republic of Korea)ELDERMET (Ireland)

Current co-chairs: G. Weinstock (WashU) & J .Peterson (NIH)

The MetaHIT approach to relation between microbes & us  

Establish a reference gene set by metagenomic & genomic sequencing of the Human GI tract microbes

Develop generic tools for profiling the GI tract microbiota genes : arrays and high throughput DNA sequencing

Use the profiling tools to search association of microbial genes and disease in Inflammatory Bowel Diseases and Obesity

Bioinformatics overlays all activities

Carry out function analysis to go from associations to mechanisms

The partners

Most important – the people, a stellar team!

• Thirteen European, one Chinese Institutions 

• Eight countries, two continents

• Nine public and four private Institutions

MetaHIT budget : 21.2 M €EC Contribution: 11.4 M €

100 scientists

Starting date: March, 2008Duration: 4 years

Illumina‐based intestinal bacteria reference gene set

Qin et al., 2010

Illumina sequencing

Samples 124 individuals(85 Danes, 39 Spaniards)

Library type15 samples 200bp

109 samples

140bp350bp

Sequencing type Paired-end (PE) sequencing

Read length (bp) 45 b (15 samples)75 b (109 samples)

Sequences per sample 31million ±0.5 million

In total, ~0.58 Terabase of sequence

Contigs

Genes

The pipeline for Human intestinal microbial gene catalog 

BGI, Wang Jun et al.

The contig set• SOAPdenovo (de Bruijn graph‐based tool)

• Removal of short contigs (<500bp) 

• Removal of redundancy

Total Size Number N50 Size N90 Size Max. Length

10.3 Gb 6.6 Million 2.2 kb 0.7 kb 237.6 kb

• Assembly error rate low: 14/Mb • Comparable to 454 (Newbler): 20/Mb

BGI, Wang Jun et al.

Representation of the human gut microbiome in the contig set

Sequences from three studies were mapped on the contig set• 124 Europeans (0.58Tb Illumina) • 18 US (1.83 Gb 454 Roche)• 13 Japanese (0.79 Gb Sanger)

The contig set represents well the whole human metagenome

BGI, Wang Jun et al.

The  gene set

Metagene prediction on the contigs: 

• 14 million Open Reading FramesRemoval of redundancy : 

≥ 95 %  nucleotide identity over at least 90 % of the length of the shorter ORF

• 3.3 million ORFs 

150 times human gene complement

The  gene set is almost complete

>85% of prevalent genes of the cohort are included in the reference set (by the incidence‐based coverage richness estimator, ICE)Genes are included if present at frequency of >7x 10-7

Human intestinal microbial genes are largely shared in the cohort 

Each individual has ~540 000 prevalent genes, on average 

40 % of an individual’s genes are shared with at least 50 % of individuals of the cohort

Deeper sequencing reveals more genes

4.5 Gb

The half’n halfrule! 

Bacterial species are also largely shared • An overwhelming fraction of the 3.3 million genes, 99%, is

of bacterial origin• A bacterial genome encodes 3,364 non‐redundant genes, 

on average

• The cohort carries at least 1000‐1150 prevalent bacterial species• Each individual carries at least 167 prevalent bacterial species

• Prevalent bacterial species must largely be shared in the cohort• They are likely to be shared in the human population

Many sequenced species are shared

Illumina reads used to identify bacterialspecies and measure their abundancein different individuals of the cohort

Individuals SpeciesAll 124  18 >90% 57>50% 75

Abundance of a bacterial speciesvaries 12‐2200 fold in individuals

We are all similar!

Functions encoded by the reference gene set

EMBL, Bork et al.

Well characterised

Well & poorly characterised

All5000 new functions (≥20 proteins)

Well characterised

Well & poorly characterised

All

EMBL, Bork et al.

Minimal metagenome – functions required in the gut ecosystem

Overall view of the minimal genome metabolic pathways (~1200 functions)

The minimal genome functions in sequenced genomes

House‐keepingGutspecific?

Ipath tool, Letunic et al. TIBS, 2008

Bacterial genes/species/communities associated to a disease ?

Take-home messages• 3.3 million prevalent human intestinal bacterial genes were identified in a cohort of 124 individuals, 150 times more than the human gene complement• The gene catalog includes most of the genes identified in the studies over three continents

• Combinations of species ( i.e. bacterial communities!) are associated to disease

After the human genomethe human metagenome!!!

Diagnostic & prognostic tests - soon- arrays, sequencing, Q-PCR; immunomarkers (?)Better treatments - next- personalized medicine

Responders/non-respondersNovel treatments – last - modulation of microbiota

Promoters (pro & pre-biotics)Inhibitors (AB-like??)

- transplantation of microbiota?

Where do these studies lead to and when?

MetaHIT ConsortiumP. Bork J. DoréJ. Raes F. GuarnerM. Arumugam J. Parkhill

O. Pedersen

Wang Jun J. Weissenbach

Qin JunjieLi Ruiquiang 

INRA,  ex GMN. PonsJ.M. Batto      E. Le Chatelier M. Almeida    P. Renault 

Acknowledgments

Thank you for your attention