TranSigner is a Python program that assigns long reads (compatible with both ONT and PacBio) to transcripts. This tool takes in a set of reads, and the assembled transcriptome. The transcriptome must be available in both fasta and gtf/gff formats. If you only have it in a gtf/gff format, you'll need to run gffread as follows:
gffread -w transcripts.fa -g genome.fa transcripts.gtf
You can install TranSigner with pip by running:
pip install transigner
or from source:
git clone https://github.com/haydenji0731/transigner transigner
cd transigner
python setup.py install
You also need minimap2 and samtools installed. You can also use the precompiled binaries as long as they are available in your PATH. If you encounter trouble during installation, please open up a Github issue.
TranSigner consists of three modules: align, prefilter, and em (short for expectation-maximization). You can check the arguments for each module by running:
transigner [module] -h
The align module takes a set of reads contained in a fastq file and align them to a transcriptome provided as a fasta file. See below:
transigner align -q reads.fastq -t transcripts.fa -d output_dir -o alignment.bam -p threads
Next, TranSigner processes the alignment results to compute compatibility scores between reads and transcripts and filter out alignments with questioning 5' and/or 3' end positions. The recommendedfilter thresholds differ by the read type:
transigner prefilter -a alignment.bam -t transcripts.fa -o output_dir --filter -tp -1 # noisy ONT direct RNA reads
transigner prefilter -a alignment.bam -t transcripts.fa -o output_dir --filter -tp -500 -fp -600 # ONT cDNA or PacBio IsoSeq
Finally, an expectation-maximization algorithm is run as follows:
transigner em -s output_dir/scores.tsv -i output_dir/ti.pkl -o output_dir --drop --use-score
By running all three modules, you'll obtain abundances.tsv and assignments.tsv files. See below for the header information for these files:
# abundances.tsv
transcript_id read_count relative_abundance
NR_024540.1 1.7149614376942495e-28 1.6757869165652874e-21
# assignments.tsv
read_id (transcript_id_1, read_fraction_1) (transcript_id_2, read_fraction_2) (transcript_id_3, read_fraction_3) ...
57ebcba2-cde0-4096-b9b9-9bdc4306cb6c (ENST00000559163, 6.640795584395568e-07) (ENST00000559884, 0.9507564850356304) (ENST00000354296, 0.04924285088481117)
You can also add the --push flag to obtain hard, 1-to-1 assignments between reads and transcripts.
Ji, H. J. and M. Pertea (2024). "Enhancing transcriptome expression quantification through accurate assignment of long RNA sequencing reads with TranSigner." bioRxiv: 2024.2004.2013.589356. [doi:https://doi.org/10.1101/2024.04.13.589356]
Pertea, G., & Pertea, M. (2020). GFF utilities: GffRead and GffCompare [version 2; peer review: 3 approved]. F1000Research, 9:304. [doi:10.12688/f1000research.23297.2]
Danecek, P., Bonfield, J. K., Liddle, J., Marshall, J., Ohan, V., Pollard, M. O., ... & Li, H. (2021). Twelve years of SAMtools and BCFtools. Gigascience, > 10(2), giab008. [doi:10.1093/gigascience/giab008]
Li, H. (2018). Minimap2: pairwise alignment for nucleotide sequences. Bioinformatics, 34:3094-3100. [doi:10.1093/bioinformatics/bty191]
Li, H. (2021). New strategies to improve minimap2 alignment accuracy. Bioinformatics, 37:4572-4574. [doi:10.1093/bioinformatics/btab705]
Lianming Du, Qin Liu, Zhenxin Fan, Jie Tang, Xiuyue Zhang, Megan Price, Bisong Yue, Kelei Zhao. Pyfastx: a robust Python package for fast random access to sequences from plain and gzipped FASTA/Q files. Briefings in Bioinformatics, 2021, 22(4):bbaa368.