Pipeline for mining microbial data from single-cell RNA sequencing (scRNA-seq) data

Brief Description:

This pipeline can help researchers obtain microbial data from existing scRNA-seq datasets without conducting additional wet-lab experiments.

Environment:

This pipeline requires a basic UNIX/Linux environment. The computational tools required in this pipeline include Kraken2, Shortread, Trimmomatic, and umi-tools.

Usage:

1. Candidate microbial reads and taxonomic profiling identification

First, install Kraken2 and download a Kraken2 database. Here, Kraken2 database is downloaded to /database_path/kraken2_database. Second, download scRNA-seq datasets. Here, CRR034524, containing two scRNA-seq datasets (CRR034524_f1.fastq.gz and CRR034524_r2.fastq.gz), is downloaded to /path/data/CRR034524. Third, run Kraken2 using the code below:

cd /path/
mkdir p15 
mkdir result
kraken2 --db /database_path/kraken2_database --threads 56 --paired ./data/CRR034524/CRR034524_f1.fastq.gz ./data/CRR034524/CRR034524_r2.fastq.gz --classified-out ./p15/p15_classified#.fastq --unclassified-out ./p15/p15_unclassified#.fastq --report ./result/CRR034524.report --output ./result/CRR034524.output

2. Barcode extraction

In this step, each candidate microbial read is given a barcode. Install umi-tools, then run umi_tools using the code below:

mkdir ./p15/filter_reads
mkdir ./p15/filter_reads/re
umi_tools whitelist --stdin ./p15/p15_classified_1.fastq --bc-pattern=CCCCCCCCCCCCCCCCNNNNNNNNNN --log2stderr > ./p15/filter_reads/whitelist.txt
umi_tools extract --bc-pattern=CCCCCCCCCCCCCCCCNNNNNNNNNNNN --stdin ./p15/p15_classified_1.fastq --stdout ./p15/filter_reads/re/p15_classified_1_extracted.fastq --read2-in ./p15/p15_classified_2.fastq --read2-out=./p15/filter_reads/re/p15_classified_2_extracted.fastq --filter-cell-barcode --whitelist=./whitelist.txt

3. Quality control

In this step, reads with low quality are removed. Install Trimmomatic (Trimmomatic is downloaded to /trimmomatic_path/trimmomatic.jar) and Shortread, then run Trimmonatic and Shortread using the code below:

java -jar /trimmomatic_path/trimmomatic.jar PE –threads 56 ./p15/filter_reads/re/p15_classified_1_extracted.fastq ./p15/filter_reads/re/p15_classified_2_extracted.fastq -baseout ./p15/filter_reads/re/p15_trimmomatic.fastq.gz SLIDINGWINDOW:4:20 MINLEN:40
Rscript ./shortread.R ./p15/filter_reads/re

4. Microbial-cell matrix construction

Run the code below:

gunzip ./p15/filter_reads/re/shortread_clean_2.fastq.gz
python ./matrix.py -input ./p15/filter_reads/re/shortread_clean_2.fastq

Finally, the output file ./p15/filter_reads/re/read_count_label.tsv is the microbal-cell matrix file.

Contact

Any questions, problems, or bugs are welcome and should be dumped to Qin Ma.

HNUmcc/IOM