Prediction of SARS-CoV-2 Variant Lineages Using the S1-Encoding Region Sequence Obtained by PacBio Single-Molecule Real-Time Sequencing

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is the causal agent of the COVID-19 pandemic that emerged in late 2019. The outbreak of variants with mutations in the region encoding the spike protein S1 sub-unit that can make them more resistant to neutralizing or monoclonal antibodies is the main point of the current monitoring. This study examines the feasibility of predicting the variant lineage and monitoring the appearance of reported mutations by sequencing only the region encoding the S1 domain by Pacific Bioscience Single Molecule Real-Time sequencing (PacBio SMRT). Using the PacBio SMRT system, we successfully sequenced 186 of the 200 samples previously sequenced with the Illumina COVIDSeq (whole genome) system. PacBio SMRT detected mutations in the S1 domain that were missed by the COVIDseq system in 27/186 samples (14.5%), due to amplification failure. These missing positions included mutations that are decisive for lineage assignation, such as G142D (n = 11), N501Y (n = 6), or E484K (n = 2). The lineage of 172/186 (92.5%) samples was accurately determined by analyzing the region encoding the S1 domain with a pipeline that uses key positions in S1. Thus, the PacBio SMRT protocol is appropriate for determining virus lineages and detecting key mutations.

DeepConsensus: Gap-Aware Sequence Transformers for Sequence Correction

Pacific BioScience (PacBio) circular consensus sequencing (CCS) generates long (10-25 kb), accurate "HiFi" reads by combining serial observations of a DNA molecule into a consensus sequence. The standard approach to consensus generation uses a hidden Markov model (pbccs). Here, we introduce DeepConsensus, which uses a unique alignment-based loss to train a gap-aware transformer-encoder (GATE) for sequence correction. Compared to pbccs, DeepConsensus reduces read errors in the same dataset by 42%. This increases the yield of PacBio HiFi reads at Q20 by 9%, at Q30 by 27%, and at Q40 by 90%. With two SMRT Cells of HG003, reads from DeepConsensus improve hifiasm assembly contiguity (NG50 4.9Mb to 17.2Mb), increase gene completeness (94% to 97%), reduce false gene duplication rate (1.1% to 0.5%), improve assembly base accuracy (Q43 to Q45), and also reduce variant calling errors by 24%.

Illuminating the dark side of the human transcriptome with TAMA Iso-Seq analysis

The human transcriptome is one of the most well-annotated of the eukaryotic species. However, limitations in technology biased discovery toward protein coding spliced genes. Accurate high throughput long read RNA sequencing now has the potential to investigate genes that were previously undetectable. Using our Transcriptome Annotation by Modular Algorithms (TAMA) tool kit to analyze the Pacific Bioscience Universal Human Reference RNA Sequel II Iso-Seq dataset, we discovered thousands of potential novel genes and identified challenges in both RNA preparation and long read data processing that have major implications for transcriptome annotation.