qTeller

The MaizeGDB Python 3 version of James Schnable's RNA-seq processing pipeline and modular web interface.

For an examples of MaizeGDB's qTeller web interfaces visit https://qteller.maizegdb.org/

Overview

qTeller data types	Description
Genes in an Interval	Select a chromosome coordinate interval for a given genome to retrieve RNA/protein abundances.
Genes by Name	Paste a list of gene models of interest to retrieve their RNA/protein abundances.
Visualize Expression	Visualize RNA/protein abundances for a single gene, or compare abundances for two genes.

Data types are optimized for single-genome RNA-seq data, single-genome RNA-seq and protein abundance data, or multi-genome RNA-seq data.

Directories

Directory Name	Description
build_db	Scripts for constructing the SQLite DB.
web_interface	Public facing files that are served by the Apache Server.
qteller_python2.7	MaizeGDB Python 2.7 instance.

Environment Requirements

PHP Version 7

• No additional libraries required.

Python Version 3.6

• See python3_requirements.txt and python modules for a list of dependencies.

Apache Web Server Version 2.4

• No untypical customization is needed.

Installation

Apache / PHP

See an example of installing Apache on CentOS 8.

Python

Centos 8 comes with Python 3, which includes PIP.

To install additional libraries:

$ 'pip install -r python3_requirements.txt'

Additional Instructions

Upon successful installation of Python, PHP, and Apache, you can git clone this project into your Apache directory. The public-facing directories are located in the web_interface directory. Assuming a default Apache installation, the DocumentRoot in the httpd.conf would look like this:

DocumentRoot "/var/www/html/qTeller/web_interface"

See Adding new data on final steps for generating the DB.

Special instructions for "Genes in an Interval" php files:

(1) Drop-down menu changes for chromosome IDs must be manually edited in index_singlegenome.php, index_multigenome.php, and Protein_index.php files. These files must be edited in each to reflect the chromosome IDs of the target genome(s). For instance, maize has ten chromosomes with the nomenclature chr1, chr2, etc; Sorghum also has ten chromosomes, but the chromosome nomenclature is Chr01, Chr02, etc. The index*.php files must be edited to reflect your target genome's chromosome information:

(2) For index_multigenome.php only, the Genome Version drop-down menu must be edited to reflect the genomes from the multi-genome bed file. Note that <option value= for the Genome Version dropdown menu in the php file corresponds to the genome ID listed in Column 5 of the bed file. To see more in-depth examples of file formatting, click here.

Adding new data

The qTeller database generation script requires the following 3 files:

1. RNA-seq and/or protein abundance files
   • If it doesn't exist already, create the build_db/abundance directory (the abundance directory can be whatever name you want): $ mkdir build_db/abundance
   • Drop your fpkm_tracking files in the build_db/abundance directory. They must end with either the .fpkm_tracking file extension from a Cufflinks output, or if you are submitting RNA-seq or protein abundances with only the gene model ID (column 1) and abundance data (column 2), the file extension should be .txt .
2. GFF or bed file
   • Download a GFF file for your desired genome and place it in the build_db directory. Alternately, a bed file can be used, where the gene model ID is in column 4. A bed file is REQUIRED for multi-genome qTeller; the gene model ID is in column 4, and the genome ID is in column 5. Here is an example.
3. CSV file
   • Create a metadata file in CSV format so the script knows how to interpret the abundance files. Here is an example.
   • NOTE: The File_handle column specifies the name of the abundance file to load (minus the .fpkm_tracking or .txt file extension)

Assuming you have the required files, you can create the SQLite DB for RNA and/or protein abundance data using the following command:

$ cd build_db
$ python multigenome_build_qt_db.py <METADATA.CSV> --bed_file <BED.bed> --info_dir ./<ABUNDANCE> --dbname userdb # creates userdb

where <METADATA.CSV> is the CSV file (3), <GFF.gff3> is the GFF file (2), and is the directory where the abundance files are kept (1) as described above. This will create a userdb SQLite file.

To build the SQLite DB for single-genome data (with no protein abundances) from the included test data, download and uncompress this gff3 file, move to build_db, then run this command:

$ cd build_db
$ python build_qt_db_gene_protein.py test_singlegenome_metadata.csv --gff_file Zm-B73-REFERENCE-NAM-5.0_Zm00001eb.1.gff3 --info_dir ./test_singlegenome_fpkm --dbname singledb # creates singledb

To build the SQLite DB for single-genome data with both RNA and protein abundances from the included test data, download and uncompress this gff3 file, move to build_db, then run this command:

$ cd build_db
$ python build_qt_db_gene_protein.py test_singlegenome_metadata.csv --gff_file Zm-B73-REFERENCE-NAM-5.0_Zm00001eb.1.gff3 --info_dir ./test_protein_abundance --dbname proteindb # creates proteindb

To create the SQLite DB for multi-genome data from the included test data, run this command:

$ cd build_db
$ python multigenome_build_qt_db.py test_multigenome_metadata.csv --bed_file test_multigenome_NAM_merged_IDs.bed --info_dir ./test_multigenome_fpkm --dbname multidb # creates multidb

To see more in-depth examples of file formatting, click here.

• Fun fact: you can use the SQLite Viewer to easily look inside the DB and experiment with queries.

Finally, the last step is to move the generated singledb, proteindb, or multidb file into the web_interface directory:

$ mv singledb ../web_interface/

You should now be able to access qTeller through your browser.