NanopoReaTA is an R shiny application that integrates both preprocessing and downstream analysis pipelines for RNA sequencing data from Oxford Nanopore Technologies (ONT) into a user-friendly interface. NanopoReaTA focuses on the analysis of (direct) cDNA and RNA-sequencing (cDNA, DRS) reads and guides you through the different steps up to final visualizations of results from i.e. differential expression or gene body coverage. Furtherm 8000 ore, NanopoReaTa can be run in real-time right after starting a run via MinKNOW, the sequencing application of ONT.
Currently available analysis modules:
- Run Overview - Experiment statistics over time
- Gene-wise analysis - Gene-wise analysis of expression (gene counts, gene body coverage)
- Differential expression analysis - Differential expression and/or usage analysis of genes (DGE) and transcripts (DTE + DTU)
| Hardware | Min. number |
|---|---|
| RAM | 64GB |
| Threads | 12 |
| Biological input | |
| Total number of samples | 4 |
| Number of conditions | 2 |
| Min. number of samples per condition | 2 |
NOTE: All paths selected by NanopoReaTA should not contain any spaces in their names. Paths should always be named with underscores "_" instead of spaces " ". (e.g "Linux data" -> "Linux_data")
- Open a bash shell Ctrl + Alt + T.
- Type the following command to install docker and pull the docker image:
sudo apt-get install -y docker.io
sudo docker pull stegiopast/nanoporeata:referencesNOTE: With the docker image tag "references", all human and mouse reference files needed for NanopoReaTA will be automatically downloaded from GENCODE (~36 GB) and saved in root. If the root directory has limited space (< 40 GB), please use the tag "no_reference" as following and download the reference files as described below:
sudo apt-get install -y docker.io
sudo docker pull stegiopast/nanoporeata:no_reference- Once the docker image is pulled, the container can be run with the following command:
With references:
sudo docker run -it -p 8080:8080 -v /:/NanopoReaTA_linux_docker stegiopast/nanoporeata:referencesOr without references:
sudo docker run -it -p 8080:8080 -v /:/NanopoReaTA_linux_docker stegiopast/nanoporeata:no_referenceThe docker container setup will be finished when the following line occurs:
Listening on http://0.0.0.0:8080-
You can now navigate to a browser of your choice on your local machine and type in the following URL: http://localhost:8080/
-
NanopoReaTA should now appear on the browser window.
NOTE: If a new docker version is available, please remove the previous docker image first, before pulling the new version of NanopoReaTA.!
With references:
docker rmi -f stegiopast/nanoporeata:referencesdocker rmi -f stegiopast/nanoporeata:no_referenceFor a successfull usage on Windows sequencing output and output of NanopoReaTa have to be stored on the same hard-drive.
You will need one of the latest wsl systems on your computer.
-
Download docker desktop: https://www.docker.com/products/docker-desktop/
-
Install Docker application. We tested docker windows with hyper-v, which means one should not set the tick for WSL2 when installing docker-desktop. Restart the computer after docker-desktop installation.
-
Start docker desktop application. In order to use docker applications on windows, docker desktop has to run in the background.
-
When the docker application is opened the user should navigate to the settings of docker-desktop. By clicking on the Ressources tab of the docker-desktop settings window, you can define the RAM, CPU and memory that should be assigned to the docker VM. It is important to change these settings accordingly to reach optimal performance. Click on Apply & Restart when you accept the settings.
-
Open power shell as administrator via search. (Start -> Search -> right click -> Open as administrator)
-
Pull the docker image
docker pull stegiopast/nanoporeata:references- Start the docker image and mount the local system under a user-specified location; like "c:/".
docker run -it -p 8080:8080 -v c:/:/NanopoReaTA_windows_docker stegiopast/nanoporeata:references
NOTE: With the docker image tag "references", all human and mouse reference files needed for NanopoReaTA will be automatically downloaded from GENCODE (~36 GB) and saved in root. If the root directory has limited space (< 40 GB), please use the tag "no_reference" as following and download the reference files as described below:
docker pull stegiopast/nanoporeata:no_reference
docker run -it -p 8080:8080 -v c:/:/NanopoReaTA_windows_docker stegiopast/nanoporeata:no_reference*NOTE: Be aware that you can manually start the application also via docker-desktop under the images tab. You just have to press the play button on the listed docker image after the pull command mentioned above and type in the required special information. Please choose a port (e.g. 8080), define a volume (e.g. c:/) and mount it on /NanopoReaTA_windows_docker. Only if these steps are executed NanopoReaTA can operate successfully on your local data. Press the start button after insertion of additional data and navigate to the container tab on docker-desktop. You can open the container information by clicking on the three spots at the container bar. By investigating the logs you can observe the running container's processes.
The docker container setup will be finished when the following line occurs: Listening on http://0.0.0.0:8080
-
You can now navigate to a browser of your choice on your local machine and type in the following URL: http://localhost:8080/
-
NanopoReaTA should now appear on the browser window.
NOTE: If you pulled the docker image with included references, the data can be found in the "/Reference_data" folder of the docker and the following section can be skipped. Mouse and Human data from GENCODE (https://www.gencodegenes.org) are included.
The required genome and annotation files for the organism of interest can be downloaded from the GENCODE database (https://www.gencodegenes.org/), since the syntax of NanopoReaTA is suited to the respective standards. Mouse reference data can be obtained at GENCODE database: GRCm39 Release M27 (https://www.gencodegenes.org/mouse/release_M27.html). Human reference data can be obtained at GENCODE database: GRCh38.primary_assembly v40 (https://www.gencodegenes.org/human/release_40.html). BED files of the respective genome versions can be downloaded from RSeQC: (https://sourceforge.net/projects/rseqc/files/BED).
The following files need to be downloaded:
| Functionality | Datatype | Comments |
|---|---|---|
| Reference Genome | .fastq | Use primary assembly |
| Reference Transcriptome | .fastq | Use cdna files |
| GTF file | .gtf | Use the version that fits reference genome and transcriptome |
| BED file | .bed | BED files are available on (https://sourceforge.net/projects/rseqc/files/BED) |
Before running/exploiting real experiments with NanopoReaTA, we highly recommend to test the app, first (see Testing for more information). NanopoReaTA operates with a backend preprocessing pipeline based on nextflow [3] and multiple R and python based scripts for downstream analyses. All results are visualized within the R shiny [2] based frontend.
When the application is started, the welcome page is shown and contains a Start analysis button as well as the NanopoReaTA manual.
After pushing the Start NanopoReaTA button the user is linked to a metadata creator page. If the user already created his or her own metadata file, he or her can skip this step and provide the path to the file in the configuation page. Please be aware of the format described below (tab-separated). If no metadata file is available, the user should enter the samples, conditions, and replicates of the running sequencing experiment and must then download the self-created metadata file. If samples are barcoded the samples must be named after their barcodes (barcode01-barcode96) - meaning the folder names that MinKNOW automatically creates must match. Once the self-created metadata is downloaded it can be locally renamed and moved. By clicking the blue arrow on the bottom right of the page the configuration of the processing can be initiated.
| Samples | Condition | Replicate | Custom |
|---|---|---|---|
| Sample1 | Cond1 | R1 | male |
| Sample2 | Cond2 | R1 | female |
| Sample3 | Cond1 | R1 | female |
| Sample4 | Cond2 | R1 | male |
The user will be linked to the configuration page and has to select required files and folders or upload an already existing configuration file in yaml format from previous NanopoReaTA runs (Please check example_conf_files for correct parameter naming). Under linux, the host system is mounted on /NanopoReaTA_linux_docker and under windows it is mounted on /NanopoReaTA_windows_docker. To locate files on your computer, you must navigate into the respective mount folder first.
The following parameters have to be set by the user: *Directory inputs needs an "/" at the end. Please make sure to let them end with an "/" character if manually written in the config.yaml file.
| Parameter | Datatype | Comments |
|---|---|---|
| Number of threads | integer | Can be adjusted on the interactive scale bar |
| Run preprocessing | bool | Select yes or no (Yes if you want to run the Nextflow pipeline, no if your data is already preprocessed by NanopoReaTA) |
| Barcoded | bool | Select yes or no (Yes if your dataset is multiplexed, no if it is not multiplexed) |
| Path to main directory | string | Please insert the experiment directory created by MinKnow when the sequencing is started; |
| Path to a metadata/description file | string | Please insert the filepath to the created metadata file |
| Path to Reference genome file | string | Please insert the filepath to the reference genome - see below if docker "references" mode |
| Path to Reference transcriptome file | string | Please insert the filepath to the reference transcriptome |
| GTF annotation file | string | Please insert the filepath to the GTF file |
| BED annotation file | string | Please insert the filepath to the BED file |
| Output directory | string | Please insert the output directory file (Note that the ouput directory should already exist as an empty instance) |
After all configurations are set, the configurations will be saved as config.yaml in the defined output folder. ("output directory")
The selection can be confirmed by clicking the button at the bottom right of the tab =>.
In this tab the metadata file is shown and the user can check whether all information are loaded correctly. For pairwise comparison, the user needs to select one of the columns containing two conditions that will be compared in further analyses. For visualizations, the user can change the color-coding for each condition here. The second selected condition will be used as reference/base level of the subsequent comparisons.
By clicking on Settings overview, the user will be forwarded to the final configuration overview.
The input configurations can be finally checked by the user. If the parameters are correct, the user can start the preprocessing by clicking the Start button. Otherwise the user can rearrange the settings by going back to the configuration tab.
-
Start NanopoReaTA's UI and select Run Preprocessing - Yes at the Configuration Page to start the backend preprocessing pipeline within the app. Press the "Start" button to initiate the preprocessing. One can keep track of the running nextflow pipeline within the index.log and error.log files created in the user-defined output folder by executing
tail -f /path/to/output/dir/index.login a terminal window. -
For visualization of NanopoReaTA preprocessed results only, start NanopoReaTA's UI, select Preprocessing - No and set the respective output folder created by NanopoReaTA at the Configuration Page, before pressing the "Start" butto 8000 n. Preprocessing will not be executed.
The Run Overview tab shows the number of mapped reads and gene counts and visualizes the sample- and group-wise read length distribution and gene expression variability per preprocessing iteration. Additionally, the time each tool needs in each iteration is shown. All information is constantly updating when preprocessing is running.
The table in this tab shows the number of mapped genes (minimap2 [2]), gene counts (featureCounts [3]) and transcriptome counts (salmon [6]). The counts are provided for each sample, respectively.
One can see the read length distributions for respective samples and conditions. The read length information is extracted directly from the fastq files (MinKNOW-defined passed reads only).
On the left side the number of genes detected is plotted per iteration for samples and selected conditions, respectively. The information is extracted from the output count table of featureCounts.
On the right side the deviation of relative gene abundancy compared to the last iteration is plotted. This is a measure for the change of gene abundancy variability within a single sample. The latter allows an assumption whether relative abundancies have stabilized throughout the ongoing sequencing.
This plot visualizes the run time for each tool running during the preprocessing. Thus, one is able to estimate the runtime for an update in the next iteration. Transcriptome and genome related steps run in parallel to optimize performance. All processing steps run in an additive manner to avoid redundant computational operations.
For the following analytical steps the preprocessing should be temporarily stopped and the completion of the running iteration should be awaited. The stop preprocessing button on the left side causes the pipeline to stop after each completed preprocessing iteration. Subsequently, all the analytical steps of interest can be performed. The resume preprocessing button causes the pipeline to continue once all the analytical steps of interest are performed. Don't forget to "resume" it! ;)
In the Gene-wise anaylsis tab, one is able to explore the expression levels and the gene body coverage of particular genes of interest. Be aware that at least two samples per condition have to be considered in order to use this functionality.
The table on the lefthand side lists all the genes annotated in the loaded GTF file. One can search and select several genes of interest via click on the table entry. Once a gene is selected it will occur on the table at the right hand side. By clicking the submit genes button, the analysis will start. A median of ratio normalization via DESeq2 [4] will be performed and the user can plot the raw and normalized counts per condition as Dot-, Violin- or Boxplot.
Here, one gene can be selected for gene body coverage analysis each time. The gene selection functions similar as in Gene counts. After the gene selection is submitted, the percentage of coverage for a gene divided into 100 percentiles is shown sample- and group-wise (=mean). The calculation is based on the RSeQC script for gene body coverage analysis (https://rseqc.sourceforge.net/) [7].
In the Differential Expression Analysis tab, the user can run three different analyses: Differential Gene Expression (DGE), Differential Transcript Expression (DTE) and Differential Transcript Usage (DTU) by clicking the respective button. Note that these analyses do not update automatically when processing will be started again and new data is generated. That means that after stopping the preprocessing pipeline again, the analyses buttons need to be pressed to analyse latest input files (like counts files). Once the analysis is completed, the user will be linked to the respective analysis output tab (may take a few minutes).
Differential gene expression analysis will be performed and the following visualizations are shown:
- A table of all differentially expressed genes
- PCA analysis
- Volcano plot (DGE)
- Sample-2-Sample plot
- Heatmap of the top 20 differentially expressed genes based on p-adjusted
Differential transcript expression analysis will be performed and the following visualizations are shown:
- A table of all differentially expressed transcripts
- PCA analysis
- Volcano plot (DTE)
- Sample-2-Sample plot
- Heatmap of the top 20 differentially expressed transcripts based on p-adjusted
Differential transcript usage analysis will be performed with DEXSeq and DRIMSeq. The following visualizations are shown:
- Tables of all differentially used transcripts' analysis results (Results of DEXSeq and DRIMSeq)
- General: To allow a general DTU overview, the log2FoldChange values from DEXSeq are plotted against the adjusted p-value (Volcano plot)
- Gene specific: One can select a gene to show the transcripts abundances within a gene of interest as boxplots per condition based on DRIMSeq's output.
We provide a dataset of cDNA extracted from 2 samples of HEK293 and 2 samples of HeLa cells.
Test data is available in a seafile repository https://seafile.rlp.net/d/7a99b8b210e44eb9b70a/. The output folder structure of MinKnow is kept intact with this plattform, which enables users to directly test the functionality of NanopoReaTA.
Test data is additionally available on the ENA (European Nucleotide Archive) with the project number PRJEB61670: https://www.ebi.ac.uk/ena/browser/view/PRJEB61670. Please note that one will have to reconstruct the folder structure of the MinKnow output using barcoded samples when dowloading files from ENA to use NanopoReaTA properly.
Examples: Experiment_folder/Sample_folder/Identifier/fastq_pass/barcodeXY/run_xyz_999.fastq with (barcode01-barcode04). Barcode01 + barcode02 are HEK293 cDNA samples and barcode03-04 are HeLa cDNA samples.
The metadata file can be found under https://github.com/AnWiercze/NanopoReaTA/blob/master/example_conf_files/example_metadata.txt.
A pre-print of this tool is published on bioRXiv:
Anna Wierczeiko, Stefan Pastore, Stefan Mündnich, Mark Helm, Tamer Butto, Susanne Gerber (2022). NanopoReaTA: a user-friendly tool for nanopore-seq real-time transcriptional analysis. bioRxiv 2022.12.13.520220; doi: https://doi.org/10.1101/2022.12.13.520220
[1] Anders, S., Reyes, A., & Huber, W. (2012). Detecting differential usage of exons from RNA-seq data. Genome Research, 22(10), 2008–2017. https://doi.org/10.1101/GR.133744.111
[2] Li, H. (2018). Minimap2: pairwise alignment for nucleotide sequences. Bioinformatics, 34(18), 3094–3100. https://doi.org/10.1093/BIOINFORMATICS/BTY191
[3] Liao, Y., Smyth, G. K., & Shi, W. (2014). featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics, 30(7), 923–930. https://doi.org/10.1093/BIOINFORMATICS/BTT656
[4] Love, M. I., Huber, W., & Anders, S. (2014). Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biology, 15(12), 1–21. https://doi.org/10.1186/S13059-014-0550-8/FIGURES/9
[5] Nowicka M, Robinson MD (2016). “DRIMSeq: a Dirichlet-multinomial framework for multivariate count outcomes in genomics [version 2; referees: 2 approved].” F1000Research, 5(1356). doi: 10.12688/f1000research.8900.2, https://f1000research.com/articles/5-1356/v2.
[6] Patro, R., Duggal, G., Love, M. I., Irizarry, R. A., & Kingsford, C. (2017). Salmon provides fast and bias-aware quantification of transcript expression. Nature Methods 2017 14:4, 14(4), 417–419. https://doi.org/10.1038/nmeth.4197
[7] Wang, L., Wang, S., & Li, W. (2012). RSeQC: quality control of RNA-seq experiments. Bioinformatics, 28(16), 2184–2185. https://doi.org/10.1093/BIOINFORMATICS/BTS356
Please open an issue if you encounter any issues/troubles. However, please go over the previous issues (including closed issues) before opening a new issue, as your same exact question might have been already answered previously. Thank you!