🌟 MAPIT-seq Pipeline

👤 Author: Gang Xie, PhD candidate

🏫 Affiliation: PKU-THU-NIBS Joint Graduate Program, Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, China

📧 Email: gangx1e@stu.pku.edu.cn

📅 Date: July 25th, 2025

✅ Version: 1.7

---

MAPIT-seq (Modification Added to RNA-binding Protein Interacting Transcript Sequencing) enables the identification RNA-binding protein (RBP) target transcripts by detecting adjacent RNA editing events introduced by both hADAR2dd and rAPOBEC1. This approach allows for the simultaneous profiling of RBP–RNA interactions and the transcriptome in tissues and single cells. The MAPIT-seq pipeline is designed to detect RNA editing events from hADAR2dd and rAPOBEC1, identify RBP targets, pinpoint high-confidence RBP-binding regions and de novo discover RBP-binding motifs.

📍 Cite us

If you found this pipeline useful in your work, please cite our paper:

Cheng, Q.-X.#, Xie, G.#, Zhang, X., Wang, J., Ding, S., Wu, Y.-X., Shi, M., Duan, F.-F., Wan, Z.-L., Wei, J.-J., Xiao, J., Wang, Y. Co-profiling of transcriptome and in situ RNA-protein interactions in single cells and tissues. Nature Methods (2025). https://doi.org/10.1038/s41592-025-02774-4

🔨 Install

All softwares in our pipeline are available in conda. We recommend you to install these softwares via conda:

git clone https://github.com/WangLabPKU/MAPIT-seq
cd MAPIT-seq
conda env create -f env_specific.yml # or use env.yml, more flexible but slower 
wget https://hgdownload.soe.ucsc.edu/admin/exe/linux.x86_64.v479/bedGraphToBigWig
chmod +x Mapit src/MAPIT-seq.sh bedGraphToBigWig

Optional: Add Mapit to your conda environment PATH

/home/gangx/apps/Mapit-seq needs to be replaced with your MAPIT-seq path.

conda activate Mapit-seq
mkdir -p $CONDA_PREFIX/etc/conda/activate.d
mkdir -p $CONDA_PREFIX/etc/conda/deactivate.d

cat <<EOF > $CONDA_PREFIX/etc/conda/activate.d/activate-mapit.sh
#!/usr/bin/env sh
export PATH="\$PATH:/home/gangx/apps/Mapit-seq"
EOF

cat <<EOF > $CONDA_PREFIX/etc/conda/deactivate.d/deactivate-mapit.sh
#!/usr/bin/env sh
export PATH=\$(echo "\$PATH" | tr ':' '\\n' | grep -v '^/home/gangx/apps/Mapit-seq\$' | paste -sd:)
EOF

chmod +x $CONDA_PREFIX/etc/conda/activate.d/activate-mapit.sh $CONDA_PREFIX/etc/conda/deactivate.d/deactivate-mapit.sh

Install MAPIT dependencies

REDItools2

cd ..
git clone https://github.com/BioinfoUNIBA/REDItools2 
conda install mpi4py -c bioconda -c conda-forge
cd REDItools2
pip install -r requirements.txt

‼️ Modifications required for REDItools2 scripts ‼️

Please apply the following changes before running the pipeline:

• Add from functools import reduce at line 17 in src/cineca/parallel_reditools.py
• Replace line 573 in src/cineca/parallel_reditools.py with:
  keys = list(chromosomes.keys())
• In src/cineca/reditools.py, replace sys.maxint (line 817) with sys.maxsize
• In src/cineca/reditools.py, replace "w" (line 912) with "wt"

SAILOR and FLARE

cd ..
git clone https://github.com/YeoLab/FLARE
conda install snakemake -c bioconda -c conda-forge
pip install deeptools gffutils pyfaidx Bio

🔧Configuration

Prepare these files beforehand:

• Reference genome FASTA

• Annotation files (.gtf, .gff3)

• RepeatMasker annotation

• Known SNP datasets, split by chromosome (e.g. dbSNP, 1000 Genomes, EVS/EVA)

All configuration is stored in conf/GenomeVersion.json, e.g. conf/GRCh38.json.

Download Reference Sequence and Annotation

• Abundant RNA (rRNA, tRNA, mt-rRNA and mt-tRNA) sequence (provided in ref fold) and create the index by BWA-MEM.

• Reference genome downloaded from Gencode and UCSC

Below is an example using the human genome (GRCh38):

cd "your_ref_path"
wget https://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_40/GRCh38.p13.genome.fa.gz
wget https://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_40/gencode.v40.chr_patch_hapl_scaff.annotation.gtf.gz
wget https://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_40/gencode.v40.chr_patch_hapl_scaff.annotation.gff3.gz

# RepeatMasker
wget http://hgdownload.cse.ucsc.edu/goldenPath/hg38/database/rmsk.txt.gz  # for mouse, just replace hg38 to mm10/mm39
gzip -d *


# ERCC spike-in
wget https://assets.thermofisher.cn/TFS-Assets/LSG/manuals/ERCC92.zip
unzip ERCC92.zip

🔺 We recommend retaining only autosomal, allosomal (sex chromosomes), and mitochondrial sequences—i.e., entries with the "chr" prefix—in both FASTA and GTF/GFF annotation files.

Download known variants annotation: dbSNP, 1000Genome, EVS and EVA

dbSNP

genomeVersion=GRCh38
mkdir "your_ref_path"/${genomeVersion}_SNP
cd "your_ref_path"/${genomeVersion}_SNP

# human (hg38/GRCh38)
wget https://ftp.ncbi.nih.gov/snp/organisms/human_9606_b151_GRCh38p7/VCF/GATK/All_20180418.vcf.gz # download data in VCF/GATK fold; column 1 starts with "chr"

# mouse (mm10/GRCm38)
wget https://ftp.ncbi.nih.gov/snp/organisms/archive/mouse_10090/VCF/00-All.vcf.gz

Exome Variant Server or European Variation Archive

# human (hg38/GRCh38)
wget http://evs.gs.washington.edu/evs_bulk_data/ESP6500SI-V2-SSA137.GRCh38-liftover.snps_indels.vcf.tar.gz
tar -xvf ESP6500SI-V2-SSA137.GRCh38-liftover.snps_indels.vcf.tar.gz
for i in {1..22} X Y; do
awk '{if(substr($1, 1, 1) == "#"){print $0}else if((length($4) == 1) && (length($5) == 1)) {gsub("MT","M");{if($1 ~ "chr") print $0; else print "chr"$0 }}}' ESP6500SI-V2-SSA137.GRCh38-liftover.chr${i}.snps_indels.vcf | gzip > EVS_split_chr/chr${i}.gz
done

‼️ The original database link is currently inaccessible. You may download the dataset using the following command and then unzip it directly:

# human (hg38/GRCh38)
wget -O EVS_split_chr.zip "https://zenodo.org/records/17089899/files/EVS_split_chr.zip?download=1"
unzip EVS_split_chr.zip

# mouse (mm10/GRCm38)
wget http://ftp.ebi.ac.uk/pub/databases/eva/rs_releases/release_3/by_species/mus_musculus/GRCm38.p4/GCA_000001635.6_current_ids.vcf.gz
mkdir EVA_split_chr
zcat GCA_000001635.6_current_ids.vcf.gz | awk -v dir_SNP=EVA_split_chr '{if(substr($1, 1, 1) == "#"){print $0 > "EVA_header"}else if((length($4) == 1) && (length($5) == 1)) {gsub("MT","M");{if($1 ~ "chr") print $0 > dir_SNP"/"$1; else print "chr"$0 > dir_SNP"/chr"$1 }}}'
gzip EVA_split_chr/chr*

# mouse (mm39/GRCm39)
# Not recommended due to the lack of supporting dbSNP data for mm39 in NCBI. Consider lifting over from mm10/GRCm38 if needed.
wget http://ftp.ebi.ac.uk/pub/databases/eva/rs_releases/release_3/by_species/mus_musculus/GRCm39/GCA_000001635.9_current_ids.vcf.gz

1000 Genomes Project

# human  
wget http://ftp.1000genomes.ebi.ac.uk/vol1/ftp/data_collections/1000_genomes_project/release/20181203_biallelic_SNV/ALL.wgs.shapeit2_integrated_v1a.GRCh38.20181129.sites.vcf.gz
mkdir 1000genomes_split_chr
zcat ALL.wgs.shapeit2_integrated_v1a.GRCh38.20181129.sites.vcf.gz | awk -v dir_SNP=1000genomes_split_chr '{if(substr($1, 1, 1) == "#"){print $0 > "1000genomes_header"}else if((length($4) == 1) && (length($5) == 1)) {gsub("MT","M");{if($1 ~ "chr") print $0 > dir_SNP"/"$1; else print "chr"$0 > dir_SNP"/chr"$1 }}}'
gzip 1000genomes_split_chr/chr*

There is no any known SNP data for mouse in 1000 Genomes Project, so you could create void file in void_split_chr fold.

# mouse
chroms=($(grep '>' $genome_fasta |sed 's/>//' | awk '{print $1}' | grep 'chr' ))
for chr in ${chroms[@]}; do touch void_split_chr/${chr}; done
gzip void_split_chr/chr*

‼️ Important: Ensure all chromosome names in the *_split_chr directories begin with the "chr" prefix. If not, manually prepend "chr" to maintain naming consistency across datasets.

Create configuration file

1. MAPIT configuration

After downloading the genome sequence (.fasta), genome annotations (.gff3), RepeatMasker annotations, and known SNP datasets split by chromosome, you can run the following command to generate the configuration file. This step includes building the genome sequence index, extracting gene element annotations, and creating chromosome-split dbSNP VCF files.

Mapit config --genomeVersion GRCh38 \
             --genomeFasta "full_path"/GRCh38.p13.genome.fa \
             --ERCC "full_path"/"ERCC.fa" \
             --species human \
             --outpath "full_path" \
             --genomeAnno "full_path"/gencode.v40.chr_patch_hapl_scaff.annotation.gff3 \
             --rmsk "full_path"/rmsk.txt \
             --dbSNP "full_path"/GRCh38_SNP/All_20180418.vcf.gz \
             --1000Genomes "full_path"/GRCh38_SNP/1000genomes_split_chr \
             --EVSEVA "full_path"/GRCh38_SNP/EVS_split_chr \
             --Reditools "full_path"/Directory_of_RediTools2.0_software \
             --FLARE "full_path"/Directory_of_FLARE_software

• Options:

  -h, --help Show help message and exit

  -v GENOMEVERSION, --genomeVersion GENOMEVERSION Genome build version (e.g., GRCh38)

  -s {human,mouse}, --species {human,mouse} Species identifier (human or mouse)

  -f GENOMEFASTA, --genomeFasta GENOMEFASTA Path to genome FASTA file

  -E ERCC, --ERCC ERCC Path to ERCC spike-in FASTA file

  -a GENOMEANNO, --genomeAnno GENOMEANNO Path to genome annotation file in GFF3 format

  -r RMSK, --rmsk RMSK Path to RepeatMask annotation file downloaded from UCSC

  -o OUTPATH, --outpath OUTPATH Output directory for annotation files for Mapit

  --dbSNP DBSNP Path to dbSNP VCF file

  --1000Genomes 1000GENOMES Directory of VCF files of 1000Genomes split by chromosome

  --EVSEVA EVSEVA Directory of VCF files of EVS/EVA split by chromosome

  --hisat2Index HISAT2INDEX (Optional) Path to pre-built HISAT2 index

  --Reditools REDITOOLS Directory of RediTools2.0 software

  --FLARE FLARE Directory of FLARE software

  --overwrite Overwrite existing configuration file if present.

This command will generate a GenomeVersion.json configuration file under the conf directory of the specified output path, which is required for downstream MAPIT-seq analysis.

2. FLARE configuration

To run the FLARE pipeline, you must first generate region files that define genomic intervals for cluster identification. This can be done using the generate_regions.py script. See FLARE repository in details.

"full_path_to_FLARE"/workflow_FLARE/scripts/generate_regions.py <full/path/to/your/genome/gtf/file> <genome_name>_regions

3. Homer configuration

In an effort to make sure things are standardized for analysis, HOMER organizes promoters, genome sequences and annotation into packages. See http://homer.ucsd.edu/homer/introduction/configure.html in details.

perl configureHomer.pl -install hg38

🔷Usage

Overview

---
config:
  theme: 'base'
  themeVariables:
    primaryColor: '#25b1bbff'
    primaryTextColor: '#fff'
    primaryBorderColor: '#055516ff'
    lineColor: '#bacdf1ff'
    secondaryColor: '#006100'
    tertiaryColor: '#fff'
---
graph TD;
  A{{Clean reads}}-->B(**mapping**
  Two-round uniquely mapping);
  B-->A1
  C(**finetuning**
  Fine tune alignments);
  C-->D(**callediting**
  GATK RNA editing detection);
  D-->E(**calltargets**
  Differential editing analysis to call RBP binding targets)
  B-->A2
  F(**prepare**
  Prepare files for SAILOR);
  F-->G(**SAILOR**
  RNA editing identification);
  G-->H(**FLARE**
  Edit cluster identification);
  H-->I(**hc_cluster**
  High-confidence edit cluster identification)
  I-->J(HOMER
  RBP binding motif *de novo* identification)
  B-->A3
  K(featureCounts
  Gene expression quantification)
  K-->L(DESeq2
  Differential expressed gene analysis)
  L-->M(clusterProfiler
  Functional analysis)
subgraph A1[Transcript binding level]
  direction TB
  C
  D
  E
end
subgraph A2[RBP-binding site level]
  direction TB
  F
  G
  H
  I
  J
end
subgraph A3[Expression level]
  direction TB
  K
  L
  M
end
  N(***RBP-RNA interactome
  and transcriptome
  co-profiling***)
  A1-.->N
  A2-.->N
  A3-.->N

style A fill: #7cb8c7ff,stroke:#fff,stroke-width:2px;
style B stroke:#fff,stroke-width:2px;
style C fill: #6e9beeff,stroke:#fff,stroke-width:2px;
style D fill: #6e9beeff,stroke:#fff,stroke-width:2px;
style E fill: #6e9beeff,stroke:#fff,stroke-width:2px;
style F fill: #56caa4ff,stroke:#fff,stroke-width:2px;
style G fill: #56caa4ff,stroke:#fff,stroke-width:2px;
style H fill: #56caa4ff,stroke:#fff,stroke-width:2px;
style I fill: #56caa4ff,stroke:#fff,stroke-width:2px;
style J fill: #56caa4ff,stroke:#fff,stroke-width:2px;
style K fill: #a0ca69ff,stroke:#fff,stroke-width:2px;
style L fill: #a0ca69ff,stroke:#fff,stroke-width:2px;
style M fill: #a0ca69ff,stroke:#fff,stroke-width:2px;
style N fill: #b1a054ff,stroke:#fff,stroke-width:2px;
style A1 fill: #ffffff16,stroke:#bbb,stroke-width:2px;
style A2 fill: #ffffff16,stroke:#bbb,stroke-width:2px;
style A3 fill: #ffffff16,stroke:#bbb,stroke-width:2px;

Loading

This is a dual-omics analysis framework for MAPIT-seq data that enables the simultaneous interrogation of RBP–RNA interactions and gene expression profiles. The example codes provided demonstrate a bulk MAPIT-seq workflow focused on RBP target identification. For the gene level transcriptome analysis component, please refer to standard RNA-seq analysis procedures (external tools and workflows are typically used, e.g., featureCounts, DESeq2, and clusterProfiler).

In addition to the bulk workflow, we also provide customized pipelines for single-cell and long-read MAPIT-seq data. These can be found in the pipeline/ directory.

Mapping

The MAPIT-seq pipeline consists of five main steps. The first step combines abundant RNA filtering and read alignment:

Mapit mapping -v GRCh38 --fq R1.fq.gz  --fq2 R2.fq.gz  --rna-strandness FR -n EN -r 1  -o Mapit_result -t THREAD

• Options:

  -h, --help Show help message and exit

  -v GENOMEVERSION, --genomeVersion GENOMEVERSION Genome build version (e.g., GRCh38)

  --fq FQ FASTQ file for R1 (paired-end) or single-end reads

  --fq2 FQ2 FASTQ file for R2 (paired-end reads only)

  --rna-strandness {F,R,FR,RF} Strand-specific information for HISAT2. For single-end reads, use F or R. For paired-end reads, use either FR or RF. Detailed descriptions of this option was available in HISAT2 manual (https://ccb.jhu.edu/software/hisat2/manual.shtml)

  -n SAMPLENAME, --sampleName SAMPLENAME Sample name prefix (avoid underscores)

  -r REPLICATE, --replicate REPLICATE Replicate ID used as the second prefix in output files

  -o OUTPATH, --outpath OUTPATH Output directory (default: ./Mapit_result).

  -t THREAD, --thread THREAD Maximum threads used for computation. (default: 10)

  --ERCC Add this flag if ERCC spike-in controls were used. (default: False)

This step generates 4 subdirectories in the output path and outputs mapping .bam files.

Fine tune

Mapit finetuning -v GRCh38 -n EN -r 1 -o "output_path"/Mapit_result

• Options:

  -h, --help Show help message and exit

  -v GENOMEVERSION, --genomeVersion GENOMEVERSION Genome build version

  -n SAMPLENAME, --sampleName SAMPLENAME Sample name prefix

  -r REPLICATE, --replicate REPLICATE Replicate ID used as the second prefix in output files

  -o OUTPATH, --outpath OUTPATH Output directory (default: ./Mapit_result). Must match the path used in the mapping step.

Call editing events

Supports simultaneous analysis of multiple samples.

Mapit callediting -v GRCh38 --sampleList EN,EM,NN,NM -o "output_path"/Mapit_result --prefix RBP -e Both [-t THREAD]

• Options:

  -h, --help Show help message and exit

  -v GENOMEVERSION, --genomeVersion GENOMEVERSION Genome build version

  --sampleList SAMPLELIST Comma-separated list of sample names

  -o OUTPATH, --outpath OUTPATH Output directory (default: "./Mapit_result"). Must match the path used in the mapping and finetuning steps

  --prefix PREFIX Prefix for output directories within the result path

  -e {ADAR,APOBEC,Both}, --enzyme {ADAR,APOBEC,Both} RNA-editing enzymes used. ADAR,APOBEC: MAPIT-seq (default: Both); ADAR: HyperTRIBE or TRIBE; APOBEC: STAMP

  -t THREAD, --thread THREAD Maximum threads used for computation. (default: 10)

Call RBP targets

Performs differential RNA editing analysis. Results are saved to: ${outpath}/6-Edit_calling/${prefix}${treatSampleName}/.

Mapit calltargets -v GRCh38 -i "output_path"/Mapit_result/6-Edit_calling/RBP/RBP_Edit_GE_RPE_DATA.tsv --treatName EM --controlName EN -o "output_path"/Mapit_result -p 0.05 

• Options:

  -h, --help show this help message and exit

  -v GENOMEVERSION, --genomeVersion GENOMEVERSION Genome Build Version

  -i INPUTEDIT, --inputEdit INPUTEDIT Input table of RNA editing events. (File generated in callediting step)

  -l {transcript,gene,Both}, --level {transcript,gene,Both} Perform Differential Editing Analysis in transcript or gene(including intron) level

  --treatName TREATNAME Sample name for treatment samples in output directory

  --controlName CONTROLNAME Sample name for control samples in output directory

  -o OUTPATH, --outpath OUTPATH Output directory (default: "./Mapit_result"). Must match the path used in the mapping, finetuning and calledting steps

  -b BINSIZE, --binSize BINSIZE The length of bins that are split from transcripts. (default: 50)

  -c COVERAGE, --coverage COVERAGE Minimun coverage for valid editing sites. (default: 10)

  --supply-zero Supply zero to non-editing bins. (default: False)

  -p PVALUE, --pvalue PVALUE Maximum p-value for Poisson filter of editing sites. (default: 0.1)

  --dropTreatRep DROPTREATREP Replicate id to exclude for treatment samples in output directory. (e.g. 1,4,5)

  --dropControlRep DROPCONTROLREP Replicate id to exclude for control samples in output directory. (e.g. 1,4,5)

  --singleCell Flag for single-cell input (replicate = individual cell). (default: False)

Identifying high-confidence editing clusters and RBP binding motifs

1. Prepare files for SAILOR workflow.

Mapit prepare -v GRCh38 -n EN -r 1 -o "output_path"/Mapit_result

• Options (see finetuning step)

2. Run SAILOR workflow

Mapit SAILOR -v GRCh38 -n EN -r 1 -c coverage -o "output_path"/Mapit_result -t THREAD

• Options:

  -h, --help Show help message and exit

  -v GENOMEVERSION, --genomeVersion GENOMEVERSION Genome build version

  -n SAMPLENAME, --sampleName SAMPLENAME Sample name prefix

  -r REPLICATE, --replicate REPLICATE Replicate ID used as the second prefix in output files

  -c COVERAGE, --coverage COVERAGE Minimun coverage for valid editing sites. (default: 10)

  -o OUTPATH, --outpath OUTPATH Output directory (default: "./Mapit_result"). Must match the path used in the mapping step

  -t THREAD, --thread THREAD Maximum threads used for computation. (default: 10)

3. Run FLARE workflow to identify edit clusters

Mapit FLARE -v GRCh38 -e AG -n EN -r 1 --regions <genome_name>_regions -o "output_path"/Mapit_result -t THREAD

• Options:

  -h, --help Show help message and exit

  -v GENOMEVERSION, --genomeVersion GENOMEVERSION Genome build version

  -n SAMPLENAME, --sampleName SAMPLENAME Sample name prefix

  -r REPLICATE, --replicate REPLICATE Replicate ID used as the second prefix in output files

  --regions REGIONS FLARE configuration directory of files for regions of the genome, see FLARE configuration part

  -o OUTPATH, --outpath OUTPATH Output directory (default: "./Mapit_result"). Must match the path used in the above steps

  -t THREAD, --thread THREAD Maximum threads used for computation. (default: 10)

  -e {AG,CT}, --edittype {AG,CT} Editing types

Identify MAPIT high-confidence edit clusters

Mapit hc_cluster -v GRCh38 -n EN -o "output_path"/Mapit_result -s 150

• Options:

  -h, --help Show help message and exit

  -v GENOMEVERSION, --genomeVersion GENOMEVERSION Genome build version

  -n SAMPLENAME, --sampleName SAMPLENAME Sample name prefix

  -o OUTPATH, --outpath OUTPATH Output directory (default: "./Mapit_result"). Must match the path used in the above steps

  -s SLOPLENGTH, --sloplength SLOPLENGTH Length of High-confidence clusters expanded for up- and down-stream sides

🔎 Result structure

After completing all analysis steps, the Mapit_result/ directory will contain multiple output files and subdirectories corresponding to each processing stage. For a detailed description of the output structure, please refer to the structure.md file in the MAPIT-seq repository.

📄 License

See the LICENSE file for more details.

© 2025 WangLabPKU.