Skip to content

Latest commit

 

History

History
151 lines (100 loc) · 6.58 KB

SG-NEx_m6Anet_tutorial.md

File metadata and controls

151 lines (100 loc) · 6.58 KB
Raw

Identification of m6A with the SG-NEx samples

In this tutorial, we will perform m6A identification on the SG-NEx sample. We will use the Nanopore direct RNA-sequencing data of the A549 cell line from the SG-NEx project for this analysis. The A549 cell line was extracted from the lung tissues of a patient with lung cancer. A Python package known as m6Anet, will be used to determine m6A modification site in the cell line.
Note: This tutorial may take 15 minutes to complete.

Content

Installation

m6Anet requires Python 3.8 (or later) to run. It can be installed through PyPI using the following command. Note: This command requires pip3 to be installed.

pip install m6anet

Alternatively, it can be installed from our GitHub repository.

git clone https://github.com/GoekeLab/m6anet.git
cd m6anet
python setup.py install
cd ..  # return to the working directory

Data Access and Preparation

For this tutorial, we will use the data from the A549_directRNA_replicate6_run1 sample. The required data include:

  • feature values of the A549 sample (json file),
  • indexes of the feature values file (index file),
  • number of reads for each DRACH position (readcount file).

Download data for m6Anet

The data files can be downloaded from the SG-NEx AWS S3 bucket. Please refer to this page for a comprehensive tutorial to access the SG-NEx dataset.

To begin, we will create a directory to store the required files.

# create a directory to store the processed data for this sample
mkdir A549_directRNA_replicate6_run1

Then, we will access the S3 bucket to download the data into the created directory. There are two methods to download the data.

Method One(Recommended):

Note: The command requires you to have AWS CLI installed.

# To download the processed data
aws s3 sync --no-sign-request s3://sg-nex-data/data/processed_data/m6Anet/SGNex_A549_directRNA_replicate6_run1 ./A549_directRNA_replicate6_run1
# The directory should now contain three files. They are the json, index and readcount files.

Alternatively, we can use the URL to the S3 bucket to download the processed data.

Method Two:

cd A549_directRNA_replicate6_run1 
 
# To download the json file
wget https://sg-nex-data.s3.ap-southeast-1.amazonaws.com/data/processed_data/m6Anet/SGNex_A549_directRNA_replicate6_run1/data.json  

# To download the index file
wget https://sg-nex-data.s3.ap-southeast-1.amazonaws.com/data/processed_data/m6Anet/SGNex_A549_directRNA_replicate6_run1/data.index

# To download the readcount file
wget https://sg-nex-data.s3.ap-southeast-1.amazonaws.com/data/processed_data/m6Anet/SGNex_A549_directRNA_replicate6_run1/data.readcount

# Exit the A549 directory to return to our working directory
cd ..

You may refer to this page for the list of URLs to access the processed data for m6Anet from different samples.

Running m6Anet

Now that all the required files are ready, we can proceed to run the m6Anet analysis.

# Run m6Anet
m6anet-run_inference --input_dir ./A549_directRNA_replicate6_run1 --out_dir ./A549_directRNA_replicate6_run1 --infer_mod_rate --n_processes 4

Note: You can reduce the "n_processes" if you have lesser processes available in your machine. However, lowering the "n_processes" may increase the running time.

m6Anet returns a compressed csv file (data.result.csv.gz) that contains the modification probability and stoichiometry for each candidate site. The first few lines of the file are shown below.

result

Please refer to the m6Anet documentation for more information regarding the output file.

m6Anet also allows the pooling of replicate samples. It can be done by providing the file path to each replicate sample's folder containing the json, index and readcount files.

Example:

m6anet-run_inference --input_dir ./replicate1 ./replicate2 ./replicate3 --out_dir ./result --infer_mod_rate --n_processes 4

NOTE: This is a short tutorial to demonstrate the usage of m6Anet on the SG-NEx data. Please refer to the m6Anet documentation for the complete workflow on identifying m6A modification with m6Anet.


Access to the required files for the complete workflow

If you wish to run the complete workflow of m6Anet, you can access all the required files (fast5, fastq and bam files) from the SG-NEx S3 bucket.

# Note: Please make sure to replace the "sample_alias" with your sample name

# To download the fast5 file
aws s3 sync --no-sign-request s3://sg-nex-data/data/sequencing_data_ont/fast5/<sample_alias> ./

# To download the fastq file
aws s3 sync --no-sign-request s3://sg-nex-data/data/sequencing_data_ont/fastq/<sample_alias> ./

# To download the bam file
aws s3 sync --no-sign-request s3://sg-nex-data/data/sequencing_data_ont/bam/transcriptome/<sample_alias> ./

Meanwhile, you can also access the transcriptome fasta files and the gtf file (annotation file) using the following command.

# To download the transcriptome fasta file
aws s3 sync --no-sign-request s3://sg-nex-data/data/annotations/transcriptome_fasta ./ --exclude hg38*

# To download the gtf file
aws s3 cp --no-sign-request s3://sg-nex-data/data/annotations/gtf_file/Homo_sapiens.GRCh38.91.gtf ./

Alternatively, you can download the files from the AWS S3 browser directly.

Reference

In this tutorial, we identified m6A modification sites on the SG-NEx dataset using m6Anet. If you use m6Anet and the dataset from SG-NEx in your work, please cite the following paper.

Hendra, Christopher, et al. "Detection of m6A from direct RNA sequencing using a Multiple Instance Learning framework." bioRxiv (2021). doi: https://doi.org/10.1101/2021.09.20.461055

Chen, Ying, et al. "A systematic benchmark of Nanopore long read RNA sequencing for transcript level analysis in human cell lines." bioRxiv (2021). doi: https://doi.org/10.1101/2021.04.21.440736