🧬 rdi

RDI (Repeat Detection Index) is a Codon library used to build an index for finding repeated substrings. RDI leverages this index to peform de novo genome assembly.

0. Prerequisites

1. Codon (with Python >=3.8 interoperability enabled) and Seq:

bash -c "$(curl -fsSL https://exaloop.io/install.sh)"
export platform=$(uname -s | awk '{print tolower($0)}')-$(uname -m)
curl -L https://github.com/exaloop/seq/releases/download/v0.11.3/seq-${platform}.tar.gz \
| tar zxvf - -C ~/.codon/lib/codon/plugins

Example for Python interoperability:

export CODON_PYTHON=/usr/lib/python3.8/config-3.8-x86_64-linux-gnu/libpython3.8.so

2. CMake:

sudo apt install cmake

1. Setup

0. Clone the repo: git clone [email protected]:curtisupshall/rdi
1. Fetch submodules: make submodules
2. Install libdivsufsort: make libdivsufsort
3. Compile RDI: make rdi

2. Running the Program

RDI runs in two modes. In index mode, RDI builds a repeat detection index and writes it to disk next to your input file. In query mode, you can make queries against the index.

Index Mode

./rdi index path/to/your/file.fa

Query Mode

./rdi query -l 10 -r 6

Name	Type	Description
-h, --help	-	Help
-r, --repeats	int	Repeats
-l, --length	int	Kmer length
-i, --input	string	Path to batch query file

Batch Input File

Write each query as an ordered pair of length, then repeat count. Example:

$ cat my-file.txt:

30 20
10 50

./rdi query data/test.fa -i my-file.txt:

10 GGCCAAGGCG @712432
10 TAATCCTAGC @562644
10 TACAGGTGCC @1220261
10 TAGCTGGGTG @1831828
10 TGCGGTGGCT @1821871
10 TTTGCCATGT @266207
10 TTTTGCCATG @790679
> Ran 2 queries in 0.240325927734375 ms

3. Future Work

• Indexing strategy; particularly around perfect minimal hashing
• Parallelization

4. References

Many thanks to M. Oguzhan Kulekci for providing the indexing algorithm used in this project, as well as pseudocode, examples, and general guidance.