Syllable-PBWT

Introduction

Syllable-PBWT is a space-efficient variation of the PBWT data structure introduced by Richard Durbin. In this repository, you'll find an implementation of Syllable-Query, an algorithm for querying long haplotype matches using the lightweight Syllable-PBWT data structures.

Dependencies

• Linux
• C++ (at least GCC 5)

Installation

Clone the repository:

git clone https://github.com/ZhiGroup/Syllable-PBWT

Enter the repository and compile the code:

cd Syllable-PBWT

make

Usage Instructions

Our code allows multiple client users to query from one server program as long as they are on the same host. The server help page can be viewed by running ./server:

Flag	Description	Details
-h or --help	Show this help message
-f <FILE> or --fifo <FILE>	File path to create named pipe	For communication with clients
-i <FILE> or --input_panel <FILE>	Path to VCF input file
-b <VALUE> or --bits <VALUE>	Value of B	B must be 64 or 128, by default 128. B is the number of sites to be grouped into one syllable.
-s or --save	Will save panel data	For fast loading when rerunning this program with --load. You will be told the save file upon creation.
-l <FILE> or --load <FILE>	Path to load file	Must be the save file of a previous successful run
-g <FILE> or --gen_map <FILE>	Path to genetic map file	The genetic map will be used when the query length unit is cM. Format: Sites described by one line each with genetic location as the last tab-delimited field. To accommodate header lines, lines with a non-numeric last field will be ignored.

The client program should be run with the server's named pipe (specified with --fifo) as the only parameter (you will be told so if you run ./client with no parameters). When the server is done preprocessing, the client will be prompted with directions on how to query as well as the output format. Below are the flags used to query:

Flag	Description
-q <FILE> or --query_panel <FILE>	Path to VCF query file
-l <VALUE> or --length <VALUE>	Minimum length of desired matches
-u <STRING> or --units <STRING>	Unit of query length: sites (default), cM, or bps

Sample Use

The server and client programs are to be run from different terminals on the same host. In this example, we will use the data found in the sample_data folder. On terminal 1, run

./server -f fifo -i sample_data/sample_panel.vcf

On terminal 2, run

./client fifo

and when prompted to enter a query, type

-q sample_data/sample_query.vcf -l 255

and then you will be told the file location of your match results.

Benchmarking

For reference, the benchmark folder includes our implementation for Algorithm 3 of Sanaullah et al. against which we benchmarked Syllable-Query.

Citation

If you found our work useful in your research, please consider citing the following paper:

@article{10.1093/bioinformatics/btac734,
    author = {Wang, Victor and Naseri, Ardalan and Zhang, Shaojie and Zhi, Degui},
    title = "{Syllable-PBWT for space-efficient haplotype long-match query}",
    journal = {Bioinformatics},
    volume = {39},
    number = {1},
    pages = {btac734},
    year = {2022},
    month = {11},
    issn = {1367-4811},
    doi = {10.1093/bioinformatics/btac734},
    url = {https://doi.org/10.1093/bioinformatics/btac734},
    eprint = {https://academic.oup.com/bioinformatics/article-pdf/39/1/btac734/48448727/btac734.pdf},
}