... Just to make sure this is documented somewhere.
1. Alignment
align (contigs, graph=None): If the graph (a GFA file) is passed, its segments must exactly match the contigs.
2. Naively calling mutations, then estimating and fixing these calls' FDRs
call (contigs, bam): The contigs in the contigs file should be a subset (strict or not is fine) of the references in the BAM file. If the contigs file is a strict subset, we'll produce a BCF file only describing mutations in these contigs.
fdr estimate (contigs, bam, bcf): The contigs file and BCF header contigs should exactly match. (i guess we could allow them to not match but this should be called right after call so no reason not to) However, these contigs can be a subset of the BAM file's references.
fdr fix (bcf, fdr_info): The fdr_info file should describe all but one of the BCF header contigs; this header contig is assumed to be the decoy contig.
3. Downstream analyses
spot hot-features (bcf, gff3): All contigs described in the GFF3 file must be a subset (strict or not is fine) of the contigs described in the BCF header.
spot cold-gaps (bcf): Just consider the contigs in the BCF header.
smooth apply (contigs, bam, bcf): The contigs in the contigs file should be a subset (strict or not is fine) of the references in the BAM file, and of the contigs in the BCF header. We'll only produce reads files for contigs in the contigs file.
smooth assemble (directory of one reads file for each contig): Just assemble each reads file separately.
matrix compute (contigs, bam): The contigs in the contigs file should be a subset (strict or not is fine) of the references in the BAM file. We'll only produce mutation matrix information for contigs in the contigs file.
link nt (contigs, bam, bcf): The contigs in the contigs file should be a subset (strict or not is fine) of the references in the BAM file, and of the contigs in the BCF header. We'll only produce link graphs for contigs in the contigs file.
... Just to make sure this is documented somewhere.
1. Alignment
align(contigs, graph=None): If the graph (a GFA file) is passed, its segments must exactly match the contigs.2. Naively calling mutations, then estimating and fixing these calls' FDRs
call(contigs, bam): The contigs in the contigs file should be a subset (strict or not is fine) of the references in the BAM file. If the contigs file is a strict subset, we'll produce a BCF file only describing mutations in these contigs.fdr estimate(contigs, bam, bcf): The contigs file and BCF header contigs should exactly match. (i guess we could allow them to not match but this should be called right aftercallso no reason not to) However, these contigs can be a subset of the BAM file's references.fdr fix(bcf, fdr_info): The fdr_info file should describe all but one of the BCF header contigs; this header contig is assumed to be the decoy contig.3. Downstream analyses
spot hot-features(bcf, gff3): All contigs described in the GFF3 file must be a subset (strict or not is fine) of the contigs described in the BCF header.spot cold-gaps(bcf): Just consider the contigs in the BCF header.smooth apply(contigs, bam, bcf): The contigs in the contigs file should be a subset (strict or not is fine) of the references in the BAM file, and of the contigs in the BCF header. We'll only produce reads files for contigs in the contigs file.smooth assemble(directory of one reads file for each contig): Just assemble each reads file separately.matrix compute(contigs, bam): The contigs in the contigs file should be a subset (strict or not is fine) of the references in the BAM file. We'll only produce mutation matrix information for contigs in the contigs file.link nt(contigs, bam, bcf): The contigs in the contigs file should be a subset (strict or not is fine) of the references in the BAM file, and of the contigs in the BCF header. We'll only produce link graphs for contigs in the contigs file.utils gfa-to-fasta(graph): N/A.