This guide assumes you have your original contigs assembled into a file
original_contigs.fa and that you have mapped reads from several samples to these contigs into
Note that the assembly can be constructed using either one single sample or several (usually all) samples.
In either case, all sample reads should be mapped against the assembly to achieve the best binning performance.
The next step is then to cut contigs into smaller parts:
cut_up_fasta.py original_contigs.fa -c 10000 -o 0 --merge_last -b contigs_10K.bed > contigs_10K.fa
Generate table with coverage depth information per sample and subcontig. This step assumes the directory ‘mapping’ contains sorted and indexed bam files where each sample has been mapped against the original contigs:
concoct_coverage_table.py contigs_10K.bed mapping/Sample*.sorted.bam > coverage_table.tsv
concoct --composition_file contigs_10K.fa --coverage_file coverage_table.tsv -b concoct_output/
Merge subcontig clustering into original contig clustering:
merge_cutup_clustering.py concoct_output/clustering_gt1000.csv > concoct_output/clustering_merged.csv
Extract bins as individual FASTA:
mkdir concoct_output/fasta_bins extract_fasta_bins.py original_contigs.fa concoct_output/clustering_merged.csv --output_path concoct_output/fasta_bins