Supplementary Information
Materials and Methods: Microbial DNA extraction, microarray hybridization and data extraction
Twin pair classification
To define which twin pairs were discordant in terms weight maintenance the recommendation by Stevens et al. was used: in adults long-term weight maintenance is considered as a weight change of < 3% of body weight [1]. The average weight of the recruited subjects was 68 kg, therefore the normal variation of 3% would on average correspond to 2,0 kg for one subject. Both siblings from one twin pair may vary in weight independently. Therefore twice the normal variation, i.e. 4 kg, is used as a maximum allowed weight maintenance variation to consider twin siblings concordant in body mass. To apply this weight concordance limit to all subjects the 4 kg variation was converted into a BMI difference (using the average subject length of 1.72 m), which is 1.35 kg/m2. Twice the maximal normal weight variation was taken used to define discordance in weight maintenance, which is 2.7 kg/m2.
Microbial DNA extraction, microarray hybridization and data extraction
Fecal microbial diversity and composition was studied in detail using the Human Intestinal Tract chip (HITChip) as described previously [2]. The DNA extracts were diluted to 20 ng/μl. Bacterial 16S rRNA genes were amplified in PCR using primers T7prom-Bact-27-for (5’-TGAATTGTAATACGACTCACTATAGGGGTTTGATCC TGGCTCAG–3’) and Uni-1492-rev (5’-CGGCTACCTTGTTACGAC-3’). In vitro transcription of 500 ng of T7-16S rRNA gene amplicons for each sample were performed with the Riboprobe System (Promega, Madison, WI, USA). Subsequently, the 16S rRNA gene transcripts were labeled with CyDyes using the Post-Labeling Reactive Dye (Amersham Biosciences, Little Chalfont, UK) dissolved in 84 µl DMSO. For each sample Cy3-labelled RNA and Cy5-labelled RNA were fragmented and hybridized to the HITChip microarrays, which were produced by Agilent technologies (Agilent Technologies, Palo Alto, CA., USA). Analysis of the microarray data was performed in a custom-designed relational database which runs under the MySQL database management system (http://www.mysql.com/) [2]. Data normalization and analysis was performed using a set of R based scripts (http://www.r-project.org/) as described before [2]. As a quality threshold only those samples were accepted for which the different Cy-dye measurements displayed a Pearson’s correlation coefficient of 0.98 or higher at probe signal intensity level.
Results
Phylogenetic group conservation in twins
Monozygotic twin pairs have a more similar microbiota composition compared to random unrelated individuals. At lower phylogenetic levels the similarity compared to random unrelated subject is higher for all subgroups as well, although the degree of similarity varies extensively (Figure S1 and S2).
Core phylotypes detection
Phylotype level HITChip data shows a core microbiota can be defined (Figure S3). Different core definitions have been used in literature [3-5]. The studied individuals core phylotypes can be detected at different minimal relative abundance thresholds. The highest relative abundance thresholds that still had at least one core phylotype with in all subjects (n=80) was 0.22%. Employing quite different definitions Tap et al. do find a phylogenetic core: 66 operational taxonomic units (OTUs) with a prevalence of >50% (n = 17) account for 35.8% of the sequences of their cohort [4]. The HITChip contains probes capable of detecting 62 of the 66 core OTUs defined by Tap et al. [4], which account for 8.1% of the total signals in our twin cohort. Although the HITChip does not detect all of the OTUs defined by Tap et al., it is unlikely that the 4 missing OTUs would account more than 25% of the bacterial population in our cohort.
Bacteroidetes : Firmicutes ratio
B:F ratios were calculated by dividing the sum of signals of the level 1 Bacteroidetes group by the sum of signals of all level 1 groups belonging to the Firmicutes combined (i.e. Bacilli, all Clostridium clusters, uncultured Clostridiales, and uncultured Mollicutes). In four discordant twin pairs the Bacteroidetes to Firmicutes (B:F) ratio is found to be higher in the high BMI twin, while in four other pairs this ratio is higher in the low BMI twin. In six, concordant, control pairs one of the twins has a much higher B:F ratio compared to its respective sibling. However most twin pairs, discordant or concordant, do not show large differences in B:F ratio (Figure S4).
References
1. Stevens, J., et al., The definition of weight maintenance. 2005. 30(3): p. 391-399.
2. Rajilic-Stojanovic, M., et al., Development and application of the human intestinal tract chip, a phylogenetic microarray: analysis of universally conserved phylotypes in the abundant microbiota of young and elderly adults. Environ Microbiol, 2009.
3. Qin, J., et al., A human gut microbial gene catalogue established by metagenomic sequencing. Nature, 2010. 464(7285): p. 59-65.
4. Tap, J., et al., Towards the human intestinal microbiota phylogenetic core. Environ Microbiol, 2009. 11(10): p. 2574-84.
5. Turnbaugh, P.J., et al., A core gut microbiome in obese and lean twins. Nature, 2009. 457(7228): p. 480-4.