Skip to main content

Prediction of Genomic Islands

Project description

TreasureIsland

TreasureIsland python package is a machine learning-based Genomic Island prediction software, that uses an unsupervised representation of DNA for its prediction.

TreasureIsland is constructed from the Benbow dataset.

Dependencies :

Python >= 3.7 Tested on Linux, mac machine For mac, make sure to run: python3 -m ensurepip --upgrade

Installation:

python3 -m venv venv
source venv/bin/activate

Option1 - Use pip:

python3 -m pip install treasureisland

if treasureisland is already installed :

python3 -m pip install treasureisland --upgrade

Option2 - Locally install package:

git clone https://github.com/priyamayur/GenomicIslandPrediction.git
python3 -m pip install -e GenomicIslandPrediction

Usage:

Option1 - Run TreasureIsland directly from commandline :

Run TreasureIsland from commandline on your DNA fasta file (example DNA files provided here), output is given in csv format:

treasureisland mypath/<DNA file>.fasta [-o <output_file_path>] [-ut <upper threshold value>] 

Option2 - Run TreasureIsland from python :

The TreasureIsland package is used to find genomic island predictions which can be downloaded in csv, xlsx, txt file formats demonstrated in TreasureIsland package

Input file:

DNA sequence files in fasta format with a sequenceID.

example: >NC_002620.2 Chlamydia muridarum str. Nigg, complete sequence CACATAGCAAAACACTCAAAGTTTTTCAGCAAAAAAGCTTGTTGAAAAAATTGTTGACCGCCTGTTCACA....

Performance:

TreasureIsland takes 2-5 mins to run depending on the size of the input.

Output :

The results are shown in the following format for each genomic island:

example : NC_002620.2 1.0 130000.0 0.95597

Upper Threshold:

User also has the ability to change the upper threshold value to change the precision and recall tradeoff. upper threshold is set to 0.80 by default.

Example :

treasureisland ecoli.fasta -o gei_output -ut 0.95 Setting the upper threshold to 0.95 would increase the precision and decrease the recall performance.

Testing:

A repository containing some sample DNA files that can be downloaded to test TreasureIsland. Note : github downloads fasta file in txt format (filename.fasta.txt).

example :

treasureisland ecoli.fasta -o gei_output -ut 0.95 

Running TreasureIsland package from python:

import the Predictor class from treasureisland package:

from treasureisland.Predictor import Predictor

Instantiate the sequence with the DNA sequence file path as the argument. The DNA file used can be a fasta file.

seq = Predictor("<Path to DNA fasta file>/ecoli.fasta", "<output_file_path>") 

Optionally, change the upper threshold value.

seq.change_upper_threshold(0.9)

Get prediction data frame from sequence by running the predict method.

pred = seq.predict()

The predictions can be downloaded in text, csv, excel formats.

seq.predictions_to_csv(pred)
seq.predictions_to_excel(pred)
seq.predictions_to_text(pred)

Contact:

Feel free to contact at banerjee.p1104@gmail.com

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

treasureisland-1.1.5.tar.gz (28.8 MB view details)

Uploaded Source

Built Distribution

treasureisland-1.1.5-py3-none-any.whl (28.9 MB view details)

Uploaded Python 3

File details

Details for the file treasureisland-1.1.5.tar.gz.

File metadata

  • Download URL: treasureisland-1.1.5.tar.gz
  • Upload date:
  • Size: 28.8 MB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.0.0 CPython/3.12.1

File hashes

Hashes for treasureisland-1.1.5.tar.gz
Algorithm Hash digest
SHA256 7317ed5dc9600820c163ea698efb292c0e0c9bab9e9c89c5c9d9c86d9af39cd9
MD5 22a297cf5776b46839ed70c20ed072f6
BLAKE2b-256 5e157f330fbc91947381e0780d74a33c99998a22ac1c7cb25f821071cda225b2

See more details on using hashes here.

File details

Details for the file treasureisland-1.1.5-py3-none-any.whl.

File metadata

File hashes

Hashes for treasureisland-1.1.5-py3-none-any.whl
Algorithm Hash digest
SHA256 476dd81d3908551e52de8641e002e03677af50da32cf407a6c65904e570a7c3e
MD5 5f7462e18090cb8d2257b1506325ecf1
BLAKE2b-256 abb29cfeb77684ab5d2dcc391f5cbfd26a8ea60bde23f642dd46679489937311

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page