Additional file 1: Table S1. Metagenomic and metatranscriptomic libraries of the thermophilic cellulose-degrading consortium. Table S2. Statistics on the scaffolds obtained from de novo assembly by IDBA-UD using EE and LE metagenomes together. Table S3. Relative abundance of each genome bin and genome completeness and contamination potential estimated based on 107 ESCGs. Table S4. List of 107 HMM of ESCGs conserved in 95 of bacteria and their representation in the five Anaerolinea draft genomes for completeness estimation. Table S5. Estimation of the validate range of genome completeness and ESCG redundancy by twenty finished genomes of Chloroflexi. Table S6. List of 35 COG marker and their representation in the five Anaerolineae draft genomes for completeness confirmation. Table S7. Summary of other draft genome bins retrieved from the metagenome. Table S8. Metabolic characteristics of isolated strains of Anaerolineae. Table S9. Comparison of assembly by three different de novo assemblers. Only the EE metagenome were assembled for comparison. Table S10. Functional orthologues of genes putatively involve in electron transfer for syntrophic metabolism. Table S11. Statistic of the post-QC HTS reads of 16S rRNA gene amplicons of the attachment samples. Figure S1. Community structure of the TCF consortium showing the accumulation of Chloroflexi during long-term run of the enrichment SBR. Figure S2. (a): Table showing the in-silico DNA-DNA hybridization values (DDH) (upper diagonal) and Average Nucleotide Identity (ANI) (lower diagonal) among five curated genomes retrieved and A. thermophila UNI-1 and C. aerophila DSM14535. Genomes in the table are ordered according to the phylogenetic relationship represented by the concatenated tree based on 35 shared ESCGs (to the left of the table). The number of aligned fragments used for ANI calculation is shown in bracket under the ANI value. (b) Venn diagram showing the number of shared and unique genes between TCF-8 and TCF-13 based on KEGG orthology annotation. (c) Venn diagram showing the number of shared and unique genes among TCF-2, 5, 12 and A. thermophila UNI-1 based on KEGG orthology annotation. Figure S3. Hierarchy clustering of members of Chloroflexi based on Euclidean distance of COGs annotation of 32 available genomes of Chloroflexi phylum and five curated genomes retrieved. Finished genomes [F], permanent draft genomes [P] and draft genomes [D] were all considered to insure comprehensive functional comparison. Figure S4. Transcriptional activities of genes involve in the Glycolysis pathway (partially shown) in the five curated bins and A. thermophila UNI-1. Filled blocks in the bottom and top role respectively represent genes encoded and expressed in the corresponding genomes Blocks are filled with the same position and color as the corresponding genomes in the legend. Figure S5. Rarefaction analysis of the community attached to cellulose surface. Top and bottom sub-tables respectively represent the rarefaction curve of Archaea (top) and Bacterial (bottom) community. Figure S6. Comparison of methane (CH4) and major VFAs (acetic acid and propanoic acid) generation between Iron-supplemented (in form of Fe2O3) and control in consecutive batch tests. B1: the first batch; B3: the last consecutive batch. Figure S7. Bacterial (left) and Archaea (right) community correlation between biological replicates sampled at twelve hours. Only prevalent populations taking >1Â % of the community are considered in correlation test. Figure S8. Hydrolysis during attachment. The error bar represents the deviations between biological replicates sampled at twelve hours. Figure S9. Electrophoretogram of the 16S rRNA genes amplicons used for high-throughput sequencing. Blank represents the band of filter paper. PCR products of 30 and 48 hours was not used in the sequencing. Takara DL2000 was used as marker. Figure S10. Composition of Archaea community during attachment. Only prevalent population taking > 1Â % in the community are shown in the figure.