---
name: "delta-filter"
section: "1"
platform: "linux"
package: "mummer"
version: "3.23+dfsg-8"
release: "Debian 13 (Trixie)"
platforms: ["linux"]
url: "https://new.carta.tech/delta-filter"
---

# delta-filter(1)

read a delta alignment file from either nucmer or promer and filters the alignments

## NAME

delta-filter - read a delta alignment file from either nucmer or promer and filters the alignments

## SYNOPSIS

**delta-filter** \[options\]*\<deltafile\>*

## DESCRIPTION

- **-1**

  1-to-1 alignment allowing for rearrangements (intersection of **-r** and **-q** alignments)

- **-g**

  1-to-1 global alignment not allowing rearrangements

- **-h**

  Display help information

- **-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 (union of **-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 **-r** and **-q** options as a percent of the alignment length \[0, 100\], default 100

- **-v**

  Print the discarded alignments instead of those that pass filters

- **-b**

  Maps duplications (XOR of **-r** and **-q** alignments, one or the other but not both)

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 are output to stdout in the same delta format as the input. For multiple switches, order of operations is as follows: **-i** **-l** **-u** **-q** **-r** **-g** **-m** **-1** **-b**. If an alignment is excluded by a preceding operation, it will be ignored by the succeeding operations.

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 the **-1** and **-m** options are not required to be mutually consistent and therefore tolerate translocations, inversions, etc. In general cases, the **-m** option is the best choice, however **-1** can be handy for applications such as SNP finding which require a 1-to-1 mapping. Finally, for mapping query contigs, or sequencing reads, to a reference genome, use **-q**. 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 that 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