Clair: Exploring the limit of using a deep neural network on pileup data for germline variant calling

Luo, Ruibang; Wong, Chak-Lim; Wong, Yat-Sing; Tang, Chi-Ian; Liu, Chi-Man; Leung, Chi-Ming; Lam, Tak Wah

doi:10.1101/865782

Cited by 12 publications

(20 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In contrast, the control data had less than half the precision, recall, and F1 score ( Table 1a ). The accuracy of indel calls was lower than the accuracy of SNP calls for all datasets, which is consistent with the error profile of ONT reads as shown in previous work 21 .…”

Section: Cancer Gene Panel Enrichmentsupporting

confidence: 91%

“…We compared each call set to the Genome in a Bottle (GIAB) NA12878 small variant truth set 22 using rtg-tools to compute accuracy metrics 23 . Based on previous work 21,24 , low-complexity regions that are known to substantially reduce small variant calling accuracy were excluded. The precision, recall, and F1 scores of the UNCALLED SNP and indel calls were within one percent of the high-coverage WGS run.…”

Section: Cancer Gene Panel Enrichmentmentioning

confidence: 99%

“…We used the vcfeval command in rtg-tools 23 to compute precision, recall, and F1 scores. For the truth set we extracted all entries in the Genome in a Bottle (GIAB) NA12878 small variant truth set 22 which overlap the target genes, and removed variants which overlap low-complexity regions which negatively impact small variant detection benchmarking according to the Global Alliance for Genomics and Health (GA4GH) 21,24 . Variants in these regions were also removed from the Clair outputs.…”

Section: Small Variant Callingmentioning

confidence: 99%

“…Variants in these regions were also removed from the Clair outputs. This was done based on regions specified in previous work 21 .…”

Section: Small Variant Callingmentioning

confidence: 99%

See 3 more Smart Citations

Targeted nanopore sequencing by real-time mapping of raw electrical signal with UNCALLED

Kovaka

Fan

et al. 2020

Preprint

View full text Add to dashboard Cite

ReadUntil sequencing allows nanopore devices to selectively eject individual reads from the pore in real-time. This could enable purely computational targeted sequencing, however most mapping methods require basecalling, which is computationally intensive. Here we present UNCALLED ( github.com/skovaka/UNCALLED ), an open-source mapper that rapidly matches streaming nanopore current signals to a reference sequence. UNCALLED probabilistically considers k-mers that the signal could represent, and then prunes the candidates based on the reference encoded within an FM-index.We used UNCALLED to deplete sequencing of known bacterial genomes within a metagenomics community, enriching the remaining species by 4.46 fold. UNCALLED also enriched 148 human genes associated with hereditary cancers to 29.6x coverage using one MinION flowcell, enabling accurate detection of SNPs, indels, structural variants (SVs), and methylation in these genes. Twice as many SVs were detected compared to 50x coverage Illumina sequencing, all verified by whole-genome nanopore and PacBio HiFi sequencing.

show abstract

Section: Cancer Gene Panel Enrichmentsupporting

confidence: 91%

Section: Cancer Gene Panel Enrichmentmentioning

confidence: 99%

Section: Small Variant Callingmentioning

confidence: 99%

“…Variants in these regions were also removed from the Clair outputs. This was done based on regions specified in previous work 21 .…”

Section: Small Variant Callingmentioning

confidence: 99%

See 2 more Smart Citations

Targeted nanopore sequencing by real-time mapping of raw electrical signal with UNCALLED

Kovaka

Fan

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…Medaka version v0.10.0 (https://github.com/nanoporetech/medaka) was used with the consensus and variant subcommands. Clair callVarBam (git commit 54c7dd4) 24 was used with default ONT settings. Additional information was acquired from pysamstats version 1.1.2 (https://github.com/alimanfoo/pysamstats, pysam 0.15.2) using the variation strand (-t variation_strand) option.…”

Section: Variant Callingmentioning

confidence: 99%

High precisionNeisseria gonorrhoeaevariant and antimicrobial resistance calling from metagenomic Nanopore sequencing

Sanderson

Swann

Barker

et al. 2020

Preprint

View full text Add to dashboard Cite

The rise of antimicrobial resistant Neisseria gonorrhoeae is a significant public health concern. Against this background, rapid culture-independent diagnostics may allow targeted treatment and prevent onward transmission. We have previously shown metagenomic sequencing of urine samples from men with urethral gonorrhoea can recover near-complete N. gonorrhoeae genomes. However, disentangling the N. gonorrhoeae genome from metagenomic samples and robustly identifying antimicrobial resistance determinants from error-prone Nanopore sequencing is a substantial bioinformatics challenge.Here we demonstrate an N. gonorrhoeae diagnostic workflow for analysis of metagenomic sequencing data obtained from clinical samples using R9.4.1 Nanopore sequencing. We compared results from simulated and clinical infections with data from known reference strains and Illumina sequencing of isolates cultured from the same patients. We evaluated three Nanopore variant callers and developed a random forest classifier to filter called SNPs. Clair was the most suitable variant caller after SNP filtering. A minimum depth of 20x reads was required to confidently identify resistant determinants over the entire genome. Our findings show that metagenomic Nanopore sequencing can provide reliable diagnostic information in N. gonorrhoeae infection.

show abstract

Extension and Improvement of CRISPR-Based Technology

Zhang

Wang

Liu

2022

Crispr

View full text Add to dashboard Cite

Clair: Exploring the limit of using a deep neural network on pileup data for germline variant calling

Abstract: 13

Cited by 12 publications

References 21 publications

Targeted nanopore sequencing by real-time mapping of raw electrical signal with UNCALLED

Targeted nanopore sequencing by real-time mapping of raw electrical signal with UNCALLED

High precisionNeisseria gonorrhoeaevariant and antimicrobial resistance calling from metagenomic Nanopore sequencing

Extension and Improvement of CRISPR-Based Technology

Contact Info

Product

Resources

About