<img src="https://github.com/ashleymaeconard/TIMEOR/blob/master/app/www/timeor_logo.png" height="150">

Trajectory Inference and Mechanism Exploration with Omics in R

Author: Ashley Mae Conard

Motivation

It's about time! Click here for a quick video demonstration of the TIMEOR application, and click here for a video guide through the tutorial below. Note that there have been several great recent additions to TIMEOR that these videos do not reflect, yet the main features and the steps are still the same. Enjoy!

Analyzing time series differential gene expression and other multi-omics data is computationally laborious and full of complex choices for the various types of time series experiment analyses. The TIMEOR web-application offers a solution as an interactive and adaptive R-Shiny web interface to support reproducibility and the best tools for a given experimental design.

TIMEOR can take in either raw .fastq files or a raw pre-computed count matrix (genes by samples), and after answering six questions about your experiment, performs all analysis from quality control and differential gene expression to gene trajectory clustering and transcription factor gene regulatory network construction. TIMEOR also suggests and allows users to integrate ChIP-seq data from ENCODE to help vialidate predicted transcription factors binding to gene clusters. Through a suite of published and novel methods, TIMEOR guides the user through three stages, suggesting adaptive default methods and comparing results for multiple normalization, alignment, and differential expression (DE) methods. Those stages are Pre-processing (generating count matrix, normalizing and correcting data), Primary Analysis (DE and gene trajectory clustersing) and Secondary Analysis (enrichment, factor binding, and temporal relations).

TIMEOR’s interactive data visualizations and publication-ready figures streamline the process of time series data analysis and assist the user to design follow-up experiments. TIMEOR is available for Homo sapiens, Mus musculus, and Drosophila melanogaster. The web server is completely free, and is accessible at https://timeor.brown.edu in a partnership between the Computational Biology Core at Brown University and Harvard Medical School's DRSC TRiP Core Facility. It is also available through Docker for large dataset processing from raw .fastq files (as the TIMEOR web server has an upload file size limit of 10GB per user).

Quick Start: 3 Steps

TIMEOR accepts 2 input types: (1) raw .fastq files and SraRunTable (e.g. here) or a (2) RNA-seq time-series read count matrix (e.g. here) and metadata file (e.g. here).

1. Visit https://timeor.brown.edu.
2. For (1) in 'Example Data' (side-bar) under 'Load raw data' click the 'SraRunTable & .fastq files' button. This will guide you through the 'Set Input and Defaults, Process Raw Data' tab demo. Follow pop-ups and fill in grey boxes. See Run TIMEOR for walk-through.
3. Next, for (2) in 'Example Data' (side-bar) under 'Load count matrix' click the 'Metadata & read count file' button. This will guide you through the rest of the full method demo. Follow pop-ups and fill in grey boxes. See Run TIMEOR for full application walk-through.

Paper and Citation

Please join a number of labs using TIMEOR. Read our paper NAR paper here, and cite:

Conard, A. M., Goodman, N., Hu, Y., Perrimon, N., Singh, R., Lawrence, C., & Larschan, E. (2021). TIMEOR: a web-based tool to uncover temporal regulatory mechanisms from multi-omics data. Nucleic Acids Research, 49(W1), W641-W653.

Overview

The TIMEOR software (web and Docker) gives users a flexible and intuitive web platform with which to upload their time-series RNA-seq data and protein-DNA data, and step through the entire temporal differential gene expression and gene dynamics analysis pipeline to generate gene regulatory networks. The application is organized into three separate stages: Pre-processing, Primary Analysis, and Secondary Analysis. Click here for a quick video demonstration of the TIMEOR webserver, and click here for a video guide through the webserver tutorial (below in Run TIMEOR).

<p>   </p> <center> <img src="https://github.com/ashleymaeconard/TIMEOR/blob/master/app/www/timeor_steps.png" style="width:95.0%" /> </center> <p>   </p>

A. Pre-processing: Gather and configure time series RNA-seq data. The user can choose to process raw data (.fastq files) using a GEO identifier, or upload a raw count matrix (genes by samples). TIMEOR then automatically chooses from several methods with which to perform quality control, alignment, and produce a count matrix. The user can then choose between several methods to normalize and correct the data.

B. Primary Analysis: Use methods to perform differential gene expression analysis and determine gene trajectory clusters. TIMEOR provides two continuous and one categorical DE method for the user. Specifically, DE genes are determined using one or more of ImpulseDE2, Next maSigPro and/or DESeq2, depending on your answers to questions in the Pre-processing stage. The user can then compare (via Venn diagram) the DE results between methods with a previous study of their choice to determine which DE method results to use for downstream analysis. The user can toggle between methods to determine which results produce the most resonable results to pass on to the next Secondary Analysis. TIMEOR then automatically clusters and creates an interactive clustermap of the selected DE gene trajectories over time. The user can choose a different number of clusters if desired.

C. Secondary Analysis: Assess enrichment, factor binding, and temporal relations. The user can analyze the gene trajectory clusters using three categories of analysis in different tabs: Enrichment, identifies the genes and gene types that are over-represented within each cluster; Factor Binding, predicts which TFs are post-transcriptionally influencing the expression of each gene cluster using motif and ChIP-seq data; and Temporal Relations, identifies transcription factor (TF) gene regulatory network (GRN). <span style="color:#3F88DE"> Blue arrow: predicted TF to observed TF, experimentally determined interaction</span>. <span style="color:#D6678D"> Pink arrow: observed TF to observed TF, experimentally determined interaction</span>. <span style="color:#F7C144"> Yellow arrow: observed TF to observed TF, predicted interaction</span>. <span style="color:#5B8179"> Green arrow: predicted TF to observed TF, predicted interaction</span>. Network displayed in table format in app to enhance flexibility of GRN visualization.

Website

Computational Biology Core at Brown Univeristy and DRSC/TRiP Functional Genomics Resources at Harvard Medical School Partnership

TIMEOR is available online at https://timeor.brown.edu. This website is free and open to everyone!

Local Installation

To run TIMEOR on large raw data time-series RNA-seq data, or to run TIMEOR scripts individually, the users may use Docker and Docker Hub. First, the TIMEOR repository must be cloned (<a href="https://github.com/ashleymaeconard/TIMEOR.git" class="uri">https://github.com/ashleymaeconard/TIMEOR.git</a>). To use Docker, it must be installed (version 20.10.0 recommended).

Docker Hub and Docker:

1. Download contents of organism genome folder (/genomes_info/) into desired location (e.g. /Users/USERNAME/Desktop/test_folder/genomes_info/) to mount later.
   • The user is welcome to gather only the organism of interest. For example, for Drosophila melanogaster simply download /genomes_info/dme/
     • Mouse is /genomes_info/mmu/
     • Human is /genomes_info/hsa/
   • Link /genomes_info/: https://drive.google.com/drive/folders/1KEnpCOU0dQU5p1tnEy3o9l02NE0uYnpm?usp=sharing
2. Make sure contents of /genome_info/ are readable.
   • For example if using Drosophila melanogaster, in a console type chmod -R 777 /Users/USERNAME/Desktop/test_folder/genomes_info/dme/.
3. Run TIMEOR via Docker
   • On command line type
     • $ docker pull ashleymaeconard/timeor:latest
     • $ docker images
     • $ docker run -v /Users/USERNAME/Desktop/test_folder/:/srv/ -p 3838:3838 <IMAGE_ID>
4. Open TIMEOR Application is available by typing:
   • Shiny server will be running on port 3838. Thus, in a browser visit localhost:3838.

Or, build Docker image

NOTE: This could take a while. Please follow these commands:

1. $ cd /PATH/TO/TIMEOR/
2. Build Docker image in TIMEOR directory:
   • $ docker build -t timeor_env .
3. Fo