Emending Alignments of Spliced Transcript Reads (EASTR)

Documentation: https://ccb.jhu.edu/eastr

(\(\
(-.-)
o('')('')

---

EASTR is a tool for detecting and eliminating spuriously spliced alignments in RNA-seq datasets. It improves the accuracy of transcriptome assembly by identifying and removing misaligned spliced alignments. The tool can process GTF, BED, and BAM files as input. EASTR can be applied to any RNA-seq dataset regardless of the alignment software used.

Dependencies

Required:

• bowtie2 >= 2.5.2
• samtools >= 1.19
• Python 3.10 or newer

Optional for testing:

• gffread >= 0.12.7
• sra-tools >= 3.0.1

Getting Started

Install using conda (Recommended)

Installing with conda gives you eastr and all of the required (bowtie2, samtools) and optional (gffread, sra-tools) dependencies.

To install from the bioconda channel use the following command:

conda install bioconda::eastr

Install using pip

You can install the EASTR package directly from PyPi but you will need to ensure that all required dependencies (bowtie2 and samtools) have been installed and are in your $PATH environment variable.

To install from pip use the following command:

# python3 -m venv .venv # (OPTIONAL)
# source .venv/bin/activate # (OPTIONAL)
pip install eastr

Required Arguments

NOTE: Only one of the below input options (GTF, BED, or BAM) should be provided.

• --gtf : Input GTF file containing transcript annotations
• --bed : Input BED file with intron coordinates
• --bam : Input BAM file or a TXT file containing a list of BAM files with read alignments

Additionally, the following arguments are required:

• -r, --reference : Reference FASTA genome used in alignment
• -i, --bowtie2_index : Path to Bowtie2 index for the reference genome

Optional Arguments

• --bt2_k : Minimum number of distinct alignments found by bowtie2 for a junction to be considered spurious. Default: 10
• -o : Length of the overhang on either side of the splice junction. Default: 50
• -a : Minimum required anchor length in each of the two exons. Default: 7
• --min_duplicate_exon_length: Minimum length that a one-anchor alignment shift must meet or exceed to be considered as representing duplicated exons. It is used to differentiate between exon duplications and spurious splice alignments. Default: 27
• --min_junc_score : Minimum number of supporting spliced reads required per junction. Default: 1
• --trusted_bed : Path to a BED file path with trusted junctions, which will not be removed by EASTR.
• --verbose : Display additional information during BAM filtering, including the count of total spliced alignments and removed alignments
• --removed_alignments_bam : Write removed alignments to a BAM file
• -p : Number of parallel processes. Default: 1

Minimap2 Parameters

• -A : Matching score. Default: 3
• -B : Mismatching penalty. Default: 4
• -O : Gap open penalty. Default: [12, 32]
• -E : Gap extension penalty. Default: [2, 1]
• -k : K-mer length for alignment. Default: 3
• --scoreN : Score of a mismatch involving ambiguous bases. Default: 1
• -w : Minimizer window size. Default: 2
• -m : Discard chains with chaining score. Default: 25

Output Options

• --out_original_junctions : Write original junctions to the output file or directory
• --out_removed_junctions : Write removed junctions to the output file or directory; the default output is to the terminal
• --out_filtered_bam : Write filtered bams to the output file or directory
• --filtered_bam_suffix : Suffix added to the name of the output BAM files. Default: '_EASTR_filtered'

Other arguments

• -p : Number of parallel processes. Default: 1

Usage

The run_eastr.sh script in the tests directory demonstrates two different ways to run the EASTR pipeline: on a bamlist and on a GTF file. Below, we provide instructions for each use case.

Running EASTR on a bamlist

1. Ensure you are in the appropriate directory containing the BAM/original folder and reference files.

2. Create a list of BAM files (make sure the list contains the full paths to the BAM files):

   ls path/to/BAM/original/*.bam > bamlist.txt

3. Run the EASTR pipeline on the bamlist with the following command:

   eastr
       --bam bamlist.txt
       --reference /path/to/reference_fasta
       --bowtie2_index /path/to/bowtie2_index
       --out_filtered_bam /path/to/output/BAM/filtered  #optional
       --out_original_junctions /path/to/output/original_junctions #optional
       --out_removed_junctions /path/to/output/removed_junctions # optional
       --removed_alignments_bam #optional
       --verbose #optional
       -p 12 #optional

Running EASTR on a GTF

Run the EASTR pipeline on the GTF file with the following command:

  eastr
    --gtf /path/to/gtf_file
    --reference /path/to/reference_fasta
    --bowtie2_index /path/to/bowtie2_index
    --out_removed_junctions /path/to/output/outfile.bed # optional

Analyzing an example dataset

Note 1: Downloading FASTQ files using the get_fastq.sh script requires SRA_toolkit

Note 2: Converting the GFF reference annotation to GTF in the get_ref.sh script requires gffread

We have included a script that demonstrates the application of the EASTR pipeline to an Arabidopsis dataset featured in our study. The sra_list_arabidopsis.txt file, located in the tests directory, lists the accession IDs of the samples analyzed.

The EASTR pipeline takes BAM files as input. The run_all.sh script acquires FASTQ files, the FASTA reference and annotation, and then aligns the FASTQ files using HISAT2 to generate BAM files. These BAM files are subsequently used as input to EASTR. Additionally, EASTR can accept a GTF annotation file and output a BED file containing questionable junctions (executed in the last command of the run_eastr.sh script).

To execute the entire EASTR pipeline, which filters BAM files and identifies reference annotation errors, use the run_all.sh script found in the tests directory. This script ensures all necessary steps and subscripts are carried out in the correct order. To analyze the example dataset, follow these steps:

1. Navigate to the tests directory within the EASTR package:
2. Make sure all scripts are executable (chmod +x *sh):
3. Run the run_all.sh script.

The script will download the necessary FASTQ files, reference genome, and then perform the alignment and EASTR analysis. The output files will be generated in their respective directories within the tests folder.

When executed on 4 CPUs, the EASTR command to filter 6 BAM files completes in approximately 35 minutes, with the bulk of this time being dedicated to the filtering of BAM files (a single bam file typically takes between 15-20 minutes to filter on a single CPU). On 1 CPU, the EASTR command to identify questionable introns in an annotation takes about 30 seconds.

Citation

To cite EASTR in publications, please use the following reference:

Shinder I, Hu R, Ji HJ, Chao KH, Pertea M. EASTR: Identifying and eliminating systematic alignment errors in multi-exon genes. Nat Commun. 2023 Nov 9;14(1):7223. doi: 10.1038/s41467-023-43017-4. PMID: 37940654; PMCID: PMC10632439.