Promoter discovery using natural language processing

This repo contains code for promoter discovery and biodiversity mining using natural language processing. This effort is being pursued through three specific aims:

1. Aim 1: Develop natural processing-based model for promoter identification
2. Aim 2: Extend model to identify inducible promoter sequences
3. Aim 3: Experimentally validate promoter predictions

Promoter sequences were collected from three main databases: EPDnew, RegulonDB, DBTBS.

Directory structure

data

All data files are found in and/or will be written to data/

• data/DBTBS/
  • Contains raw data from DBTBS: Bacillus subtilis promoter database
• data/EPDnew/
  • Contains raw data from EPDnew: Eukaryote promoter database (promoter data for 15 different organisms
• data/RegulonDB/
  • Contains raw data from RegulonDB: Escherichia coli promoter database
• data/parsed_promoter_data/
  • Promoter data parsed from each database
• data/20191114promoter_identification_ML_curation
  • Manually curated information on other state-of-the-art ML models for promoter prediction

src

All code are found in and/or will be written to src/ in either notebook or script form

Notebooks:

• src/notebooks/20191125_promoter_database_parsing.ipynb
  • Notebook containing code for parsing promoter data

figs

All raw and edits figures will be writted to figs/