Alignment coding challenge

Introduction

This project is a simple DNA sequence alignment web application. Users can submit any DNA sequence(consisting of 'A','T','G' and 'C's), and the application will parse over the protein sequences in a provided list of assemblies to find an alignment. Details of the alignments are displayed. A history of query results that have been submitted is maintained.

The list of assemblies across which to conduct searches are NC_000852, NC_007346, NC_008724, NC_009899, NC_014637, NC_020104, NC_023423, NC_016072, NC_023719, and NC_027867.

Note that NC_023640 was in the 'provided list' in the challenge requirements. However, NC_023640 has been obsoleted, and replaced by NC_016072.

Implementation Notes

1. The back-end has been written in Java 8 with SpringBoot, and the front-end is in Angular 9.
2. Build and deploy can be most easily done using Docker and Docker Compose.
3. Data for each of assemblies in the 'provided list' were obtained from the corresponding NCBI Nucleotide site.
   E.g. For NC_027867, I navigated to https://www.ncbi.nlm.nih.gov/nuccore/NC_027867. I clicked on the 'Send to' link in the top right area, and downloaded a feature file for the Coding Sequences (in FASTA Nucleotide format). These files can be found in the alignment-service module, in the directory resources/coding-sequences.
4. I have written some unit/integration tests for the backend, more are needed. I have not had the time to write any unit tests for the front end.
5. A history of query requests(with results) are displayed, with the most recent at the top. On closing out of the browser, and restarting, these results will still be shown. If there were any submissions pending at the time the browser was closed, the submissions are triggered again. The history is specific to a machine. Users accessing the website from other machines will not share results.

Hosted app

This app has been hosted on DigitalOcean, and is available at http://67.205.186.166:9898.

Install and Deploy Locally

Download the code in this Github repo to a location on your machine.

With Docker

The only requirement for this mode is that Docker and Docker Compose be installed on your machine.

1. Start Docker
2. In a command line terminal, navigate to the root of the project (which contains the docker-compose.yml file)
3. If you are building the project for the first time (or want to rebuild), type
   docker-compose up --build
   If you already have a build, simply type
   docker-compose up
   
4. In a browser, navigate to http://localhost:9898 to bring up the web application interface.

Without Docker

1. Install Java 8, Angular 9, and npm on your machine.
2. In a command line terminal, navigate to the alignment-ui directory. Type the following
   > npm install
> ng serve
   This serves up the front-end.
3. The Java project is in the alignment-service directory. You can open it in an IDE of your choice, and do a 'maven clean install'. Run the main class which is imaginatively named Application.java. (There are command line means to do this too.). This starts the back-end.
4. In a browser, navigate to http://localhost:4200 to bring up the web application interface.

Application Notes

1. The query sequence can be entered in the input box at the top of the webpage.
2. The 'Sequence Match Results' section contains all queries with match details (if any). These are maintained on tab/browser reopen. This includes any pending queries.
3. The button 'Clear Results' will clear all match results.
4. When performing an alignment, the system only looks for the first match - any subsequent alignments are ignored.

Improvements

This is a very basic application, there is much to improve and optimize.

1. More unit tests should be written for both the back and front ends.
2. Currently, an exact substring match is performed. However, exactness may not be needed. Going forward, this would probably utilize a more intelligent matching algorithm.
3. Instead of reading off the assembly files in the 'resources' folder, we may want to make REST calls to NCBI instead. However, frequent calls to NCBI can experience outages. Another approach to this would be to do nightly downloads from NCBI. This data could be parsed into a database, perhaps an in-memory version.
4. Exception handling and logging needs to be implemented. ... There are many, many more ways to scale and optimize this application, but I am running out of time.

*** It's been great fun working on this app! ***