Genomic Language Model Mitigates Chimera Artifacts in Nanopore Direct RNA Sequencing

Detection of chimeric read artifacts in dRNA-seq data using DeepChopper and validation with orthogonal sequencing platforms

(b) Length distribution of predicted adapters by DeepChopper in VCaP dRNA-seq data.

(c) Distribution of relative adapter position along read length in VCaP dRNA-seq data. Grey rectangle represents a long read from 5′ to 3′. Relative position is calculated as the ratio of the length before DeepChopper predicted adapter start position to the total read length.

(d) Distribution of segments per read after trimming: 1 segment indicates 3′ end adapter trimmed, while 2 or more indicate internal adapters trimmed.

(e) Chimeric alignments (in thousands) for VCaP dRNA-seq reads processed by Dorado with and without adapter trimming, and DeepChopper. DeepChopper greatly reduces chimeric alignments not supported by direct cDNA sequencing.

(f) Distribution of base qualities from identified internal adapters by DeepChopper. Background colors indicate quality levels: green (high), yellow (medium), and red (low).

(g) BLAST-like alignment tool (BLAT) identity distribution of the internal adapter sequences mapping against human reference genome.

(h) The number of chimeric alignments (in thousands) for A549, HCT116, HepG2, K562, and MCF7 cell lines processed by Dorado with adapter trimming and DeepChopper. DeepChopper consistently reduces chimeric alignments not supported by cDNA sequencing across all cell lines.

(i) Chimeric alignments from dRNA-seq of the WTC11 cell line, evaluated for support using additional ONT and Pacific Biosciences (PacBio) sequencing data with different protocols. DeepChopper reduces unsupported chimeric alignments across all methods compared to Dorado with adapter trimming.

Characterization of dRNA-seq chimera artifacts and their impact on downstream analysis in VCaP cells

(a) The upper bar plot shows the number of chimeric connections per kilobase across chromosomes, highlighting higher chimeric activity in Chr 19 and Chr M. The lower heatmap visualizes interchromosomal connections, with intensity indicating the count of connections between different chromosomes.

(b) The bar plot shows the number of transcripts (in thousands) across different isoform classification categories. DeepChopper-processed reads result in a higher number of transcripts compared to Dorado-trimmed reads. The inset details the isoform classification scheme.

(c) Detected gene fusions from Dorado adapter-trimmed reads and DeepChopper-processed reads. Gene fusions identified from short-read RNA-seq were used to validate fusion events detected from dRNA-seq.

(d) Gene Ontology (GO) enrichment analysis of chimera artifact-affected genes, with color indicating gene count per term.

(e) Analysis of a chimeric read artifact detected as an RPS29-COX8A fusion. The schematic shows the fusion between RPS29 (Chr 14) and COX8A (Chr 11). The green plot indicates quality scores along the read, and the blue plot shows raw signal intensity (in pA). The chopped region identified by DeepChopper corresponds to a low-quality segment with low current intensity, polyA, and short open pore signals, suggesting the presence of an ONT adapter.

Benchmarking of DeepChopper predictions on chimeric read artifacts using synthetic data and dRNA-seq data generated with the SQK-RNA004 kit from the VCaP cell line

(a) Performance evaluation in a held-out test dataset showing Recall, Precision, and F1 values for DeepChopper.

(b) Number of chimeric alignments (in thousands) identified in VCaP RNA004 dRNA-seq reads processed by Dorado with and without adapter trimming, and by DeepChopper. The bar colors indicate chimeric reads supported by cDNA sequencing (red) and those lacking support (grey).

(c) Base quality scores (left) and BLAT alignment identity scores (right) for internal adapter sequences identified by DeepChopper in RNA004 dRNA-seq reads.

Analysis of dRNA-seq chimera artifacts and their genomic and transcriptomic characteristics in VCaP cells

(a) Box plot comparing gene expression levels between all expressed genes (N=19,156) and genes affected by chimera artifacts (N=7,186) in the VCaP dRNA-seq dataset. Chimera artifacts-affected genes exhibit significantly higher expression levels (p-value < 2.2 × 10^-16).

(b) Distribution of gene effective sizes for all expressed genes and genes affected by chimera artifacts, indicating that the size distributions of genes impacted by chimera artifacts are comparable to those of all expressed genes.

(c) Chromosomal distribution and interchromosomal connections from chimeric read artifacts arising from VCaP RNA004 dRNA-seq. The top bar plot shows the number of connections per kilobase for each chromosome, with higher bars indicating more frequent connections. The bottom heatmap visualizes the number of chimeric connections between chromosome pairs, with color intensity representing the connection frequency.

(d) Number of detected transcripts across different isoform categories (FSM, ISM, NIC, NNC, and Other) from DeepChopper-identified chimeric read artifacts in VCaP RNA004 dRNA-seq data. DeepChopper-corrected reads resulted in a greater number of transcripts compared to adapter-trimmed reads by Dorado across all categories.

Chimeric alignments from dRNA-seq of the F121-9 cell line (mouse), evaluated for support using additional ONT and PacBio sequencing data with different protocols. DeepChopper reduces unsupported chimeric alignments across all methods compared to Dorado with adapter trimming.

Analysis of gene fusions derived from chimeric read artifacts in dRNA-seq

(a) Circos plot depicting chromosomal connections of gene fusions resulting from chimeric read artifacts in VCaP cells. Blue lines represent inter-chromosomal fusion events, while red lines indicate intra-chromosomal fusions. The outer track displays chromosomal ideograms labeled with respective chromosome numbers.