structural_DLFA

Deep learning and structure based hypothesis generation for functional annotation

A pipeline for bringing together hypotheses from SAdLSA and AlphaFold to uncover protein structure.

---

Directories

• database -- contains Database, which is a wrapper for our sqlite3 database
• notebooks -- jupyter notebooks that demonstrate how to use some this software as well as create various visualizations; also some notebooks are for working out various issues.
• output -- CSV output from earlier SAdLSA runs
• parsers -- this package contains modules for parsing Alphafold, SAdLSA, and GenBank data.
• scripts -- scripts for ingesting data into sqlite3 database as well as lookup proteins in Genbank files.
• seq -- data saved from an earlier SAdLSA run
• snippets -- example code from which to crib
• visualizers -- code for visualizing Alphafold and SAdLSA data
• website -- support for our Missouri web site for querying and visualizing sqlite3 data.

Files

• requirements.txt -- python dependencies for database, scripts, and parsers
• Dockerfile -- for building web site container
• docker-compose.yml -- for spinning up the container

---

Spinning up web site

1. Ensure your environment is setup
   1. docker is installed

      We recommned installing via the website and not via home brew or a linux package manager.

   2. git lfs is installed
   3. you have checked out website/static/dlfa.db via git lfs
2. docker-compose up

   You will see logging messages in your terminal for all GET and POST transactions

3. Surf to https://0.0.0.0:5000

(Yes, eventually the port and host IP will change.)