A tool for assigning chromosome labels based on a reference genome. This method is intended to be more sensitive than whole-genome alignment.
No installation is needed. Just install the above dependencies and run python3 chrom_match.py
Suppose there is a "target" genome assembly with 12 chromosome-scale sequences and the goal is to assign chromosome names based on a related reference genome, also with 12 chromsomes.
python3 chrom_match.py target.fa reference.fa reference.genes.gff3
Suppose that the target and reference assemblies have additional sequences, such as unplaced contigs/scaffolds. Then, supply the sequences to be matched with -t and -r.
In all cases, the number of target and reference sequences must be the same in order to be matched.
- Write reference transcripts to a FASTA file
- Align these transcripts to the target assembly with minimap2
- Process these alignments to build a bipartite graph
- For each gene, only consider its longest (representative) transcript
- Only consider representative transcripts if they align with mapq > 10 and coverage >= 85%
- Target sequences make one set of nodes
- Reference sequences make the other set of nodes
- Edges connecting these nodes have their weight decremented by one for each shared transcript
- Compute a minimum weight full matching for the graph
- Output the matching solution in AGP format