DELTA-FILTER(1)

🌸 NAME 🌸

okay bestie, meet delta-filter — she reads a delta alignment file from either nucmer or promer and filters the alignments. that's the whole job description, and she does it flawlessly. she lives in the User Commands part of the manual, iykyk 💖

✨ SYNOPSIS ✨

delta-filter [options]<deltafile>

💅 DESCRIPTION 💅

🎀 the flags, one by one 🎀

• -1 — 1-to-1 alignment, and it allows for rearrangements (it's the intersection of the -r and -q alignments, bestie, literally the overlap).
• -g — 1-to-1 global alignment, and this one does not allow rearrangements. no shuffling. none. 🙈
• -h — display help information. she's telling you how to love her.
• -i float — set the minimum alignment identity [0, 100], default 0.
• -l int — set the minimum alignment length, default 0.
• -m — many-to-many alignment allowing for rearrangements (the union of the -r and -q alignments).
• -q — maps each position of each query to its best hit in the reference, allowing for reference overlaps.
• -r — maps each position of each reference to its best hit in the query, allowing for query overlaps.
• -u float — set the minimum alignment uniqueness, i.e. percent of the alignment matching to unique reference AND query sequence [0, 100], default 0.
• -o float — set the maximum alignment overlap for the -r and -q options as a percent of the alignment length [0, 100], default 100.
• -v — print the discarded alignments instead of the ones that pass the filters. like showing you the ones that didn't make the cut. iconic honestly.
• -b — maps duplications (XOR of the -r and -q alignments — one or the other but not both).

so here's the tea 🫖 — she reads a delta alignment file from either nucmer or promer and filters the alignments based on the command-line switches, leaving only the desired alignments, which get output to stdout in the same delta format as the input. clean. no notes.

if you're stacking multiple switches, the order of operations is literally this:

-i -l -u -q -r -g -m -1 -b

and if an alignment gets excluded by a preceding operation, it will be ignored by the succeeding operations. so it's out for good, no second chances, periodt.

🌈 the -g vs -1 vs -m situation (important) 🌈

an important distinction between the -g option and the -1 and -m options is that -g requires the alignments to be mutually consistent in their order, while -1 and -m are not required to be mutually consistent and therefore tolerate translocations, inversions, etc. she's flexible like that. 💅

in general cases, the -m option is the best choice — main character energy fr fr. however, -1 can be handy for applications such as SNP finding, which require a 1-to-1 mapping. and finally, for mapping query contigs, or sequencing reads, to a reference genome, use -q. that's her moment. ✨

the duplications printed with the -b option are -r and -q alignments that are not present in the 1-to-1 alignment. these alignments are also the difference between the -1 and -m alignments. 🦋

🦋 SEE ALSO 🦋

the -b option originates from mugsy, which provides a code copy of mummer with additional patches. the source can be found in SVN:

svn://svn.code.sf.net/p/mugsy/code/trunk

and that's the whole page, bestie — brought to you by mummer 3.23, dated April 2015. she's vintage, she's stable, she's slaying. 💖✨🎀🫶