rhagenson / swsc

swsc implements the Sliding-Window Site Characteristics method of ultraconserved elements (UCEs) detailed by 10.1093/molbev/msy069

swsc

swsc implements the Sliding-Window Site Characteristics (SWSC) method as described in https://doi.org/10.1093/molbev/msy069

Sample data taken from: PFinderUCE-SWSC-EN

Insight into Interworkings

If using this program as a blackbox, here are a few things to consider about the interworkings:

swsc Window Types

swsc consides three window types starting at v5.0.0:

• Candidate windows (size of minWin)
• Extended candidate windows (size of minWin*2 via extending minWin/2 in both directions)
• Window covering all candidates (size between minWin*candidates and UCE length)

Change minWin and candidates

The default settings for these values are provided as a rough guide to realistic values, but are not meant to be the values used for all runs.

For best results, minWin*candidates should be roughly 1/3 of the smallest UCE, indicating candidates can span the full length of the smallest UCE.

Usage

Installation

1. Install the Go language following instructions at: https://golang.org/
2. Run go get -u github.com/rhagenson/swsc/...
3. Run swsc by either:
   • Calling it directly if you added$GOPATH/bin/ to your $PATH
   • Navigating to $GOPATH/src/github.com/rhagenson/swsc/ and running go build main.go && ./swsc

Running

Both input,output, and one metric (--gc or --entropy) must be set. See swsc --help for details.

Reporting Errors

If you have found an error, or this tools does not work for you, please create an issue at: https://github.com/RHagenson/swsc/issues with details on when the error occurred, what the error states, and what was expected to occur, if known.

Input

swsc reads a single nexus file processing two blocks:

1. DATA, containing the UCE markers (unique by ID)
2. SETS, containing the UCE locations (unique by ID, with inclusive range)

Example (... denotes truncated content, see PFinderUCE-SWSC-EN for full file):

#NEXUS

BEGIN DATA;
DIMENSIONS  NTAX=10 NCHAR=5786;
FORMAT DATATYPE=DNA GAP=- MISSING=?;
MATRIX

sp1    AGAAAC...TGCAAAG
...
;

END;

BEGIN SETS;

    [loci]
    CHARSET chr_2828 = 1-376;
    CHARSET chr_4312 = 377-627;
    ...

    CHARPARTITION loci = 1:chr_2828, 2:chr_4312...;

END;

Output

swsc writes a .csv file containing the chosen characteristic for each site of the UCEs. It can also produce a .cfg for use by PartitionFinder2 by using the appropriate flag (--cfg).

Versions

A quick explanation of versions:

• v1.0.0: Does a brute force search considering all possible windows minWin and up
• v2.0.0: Uses candidate windows plus extension procedure (optimize large alignment performance)
• v3.0.0: Candidate windows plus extension while using a single reference alignment (better performance)
• v4.0.0: Candidate windows plus extension, single reference alignment, and remove redundant calculations
• v5.0.0: Multiple candidate windows plus extension and single reference alignment
• v5.1.0: Same as v5.0.0, but done in parallel for each UCE
• v6.1.0: Update CLI to allow Nexus or FASTA+UCE csv input

Use git checkout <version> to move to a particular version (and git checkout master to move to the latest untagged development version). From there you can run either go install to install the particular version in GOPATH (overwriting any previous installed version) or go build [-o <build name>] to build the version in the current directory.

Versions can give different results so I would recommend using v1.0.0 if you want the absolute best result (and have the time to wait for it to run a long, long time) or v6.1.0 with a realistic --minWin and --candidates settings (rule of thumb: minWin*candidates should be roughly 1/3 of the smallest UCE, indicating candidates can span the full length of the smallest UCE).

License

This work is licensed under the the Modified BSD License, see full terms of use in LICENSE file.