polyTailor

• Brief description
• Installation
• How to run?
• Examples
• Test dataset
• Citation
• Issues

Brief description

PolyTailor is a software to study RNA tails. It predicts:

• per-read polyA tail lengths estimates
• per-read tail heterogeneity (i.e. non-A features)

This software is meant to be used with Nano3P-seq cDNA libraries, and can work with Nano3P-seq libraries sequenced using R9 and R10 flowcells.

Installation

You'll need:

• samtools v1.19+
• minimap2 v2.28+
• IsoQuant 3.5+
• Python 3.8+ with following packaged installed: matplotlib numpy parasail pybedtools pysam pandas scipy seaborn htseq

All above can be installed using conda and pip:

conda create -c conda-forge -c bioconda -n polyTailor python=3.10 samtools minimap2 isoquant
conda activate polyTailor
pip install matplotlib numpy parasail pybedtools pysam pandas scipy seaborn htseq
mkdir -p ~/src && cd ~/src
git clone https://github.com/novoalab/polyTailor.git

• dorado v0.7.2+ and basecalling models. Here, we assume Linux x64 system. For other systems, please see dorado page .

mkdir -p ~/src && cd ~/src
wget https://cdn.oxfordnanoportal.com/software/analysis/dorado-0.7.2-linux-x64.tar.gz
tar xpfz dorado-0.7.2-linux-x64.tar.gz
echo 'export PATH=~/src/dorado-0.7.2-linux-x64/bin:$PATH' >> ~/.bashrc
source ~/.bashrc

dorado download --directory ~/src/dorado/models --model dna_r9.4.1_e8_hac@v3.3
dorado download --directory ~/src/dorado/models --model dna_r9.4.1_e8_sup@v3.3
dorado download --directory ~/src/dorado/models --model dna_r10.4.1_e8.2_400bps_hac@v5.0.0
dorado download --directory ~/src/dorado/models --model dna_r10.4.1_e8.2_400bps_sup@v5.0.0

How to run?

PolyTailor can be executed as follows (steps 3-4 are optional):

1. Basecall (and demultiplex) your reads saving the mv table in BAM file

dorado basecaller -x cuda:all --emit-moves -r MODEL [--kit-name BARCODING_KIT] pod5_dir > reads.bam

For the most accurate poly-T composition calling we recommend using the latest sup model. If barcoding --kit-name is provided, barcode will be reported in barcode column.

2. Align the reads to the genome passing dorado tags to the resulting BAM file

samtools fastq -F2304 -T mv,ts,BC reads.bam|minimap2 -y -ax splice:hq genome.fa -|samtools sort --write-index -o algs.bam

3. Annotate alternative transcript ends

src/get_transcript_ends.py --firststrand -q0 -o transcript_ends.tsv.gz -a genome.gtf -b algs.bam [algs2.bam ... algsN.bam]

4. Associate reads to transcripts

isoquant.py --complete_genedb --data_type nanopore -o isoquant -r genome.fa -g genome.gtf --stranded reverse --bam algs.bam

5. Estimate poly-T tail length and composition

src/get_pT.py -o pT.tsv.gz -b algs.bam \
  -e transcript_ends.flt.tsv.gz.bed
  -i <(zgrep -v '^#' isoquant/OUT/OUT.read_assignments.tsv.gz | cut -f1,4,6,9)

Output

PolyTailor will produce a TAB-delimited file with following columns:

1. read_id
2. barcode - detected barcode (unknown if no barcode detected)
3. mapq - mapping quality
4. filter - OK means that following filters were passed
   • read has 5' clipped part corresponding to: adapter, N3PS primer and pT (otherwise no_clip)
   • N3PS primer was detected in the 5' clipped part (otherwise no_primer)
   • the N3PS primer aligned end-to-end (otherwise not_complete)
   • pT sequence was detected between primer and aligned transcript (otherwise no_pT)
   • the pT is immediately following primer (otherwise not_continuous)
5. pt_len - estimated poly-T length.
6. per_base - helicase speed estimated from mv table (mean number of chunks per base)
7. primer_end - position of the N3PS primer end in the read sequence
8. pt_start - position of the poly-T start in the read sequence
9. before_pt - sequence before detected poly-T (terminal bases of poly-A)
10. pt_seq - poly-T sequence composition
11. transcript_end - present if read end is associated with any of predicted transcript ends
12. distance - distance from the transcript end
13. comments - additional fields passed from -i / --readids file

Note, there may be multiple comments columns, depending on provided -i / --readids file.

For isoquant example above, you'll see:

13. isoform_id
14. assignment_type
15. additional_info

Test dataset

You can find test data and example outputs in test.

Citation

If you find this work useful, please cite: Begik O*, Pryszcz LP*, Niazi AM, Valen E, Novoa EM. Nano3P-seq: a protocol to chart the coding and non-coding transcriptome at single molecule resolution. (in preparation)

Issues

If you have an issue running this code, please open a new Github issue. Please take a look at previous issues, even if closed. Thanks!