This video provides a comprehensive overview of soft clipped reads in next-generation sequencing (NGS). It defines a soft clipped read as a sequencing read that only partially aligns to a reference genome. The video explains that these reads can originate from either true biological phenomena, such as structural variants and repetitive DNA regions, or from technical artifacts. A significant focus is placed on artifacts generated during library preparation, comparing mechanical fragmentation (the "gold standard" with the fewest artifacts) against enzymatic methods. It highlights that transposon-based enzymatic kits can generate substantially more soft clipped reads than high-quality endonuclease-based kits. The mechanism for artifact formation is detailed, involving the creation of stem-loop structures at inverted DNA repeats during library prep. Finally, the video discusses the impact on data analysis—noting a higher risk for rare variant studies—and presents mitigation strategies, including careful workflow selection, pilot testing, and the use of specialized bioinformatics tools like FADE to identify and filter artifacts, thereby preventing false positive variant calls.
Understanding and Mitigating Soft Clipped Reads in NGS
(0:00:13) What is a Soft Clipped Read?: A soft clipped read is a sequencing read where a portion at the beginning or end does not align with the reference genome. Bioinformatics analysis flags these reads to later distinguish between true biological variants and technical artifacts.
(0:01:02) Biological Origins of Soft Clipped Reads: These reads can represent real genomic information, such as structural variants (insertions, deletions, translocations), or originate from inherently challenging genomic areas like low-complexity and repetitive regions.
(0:01:49) Technical Artifacts from Library Prep: A primary source of non-biological soft clipped reads is the sample preparation process. The choice of DNA fragmentation method during library preparation significantly impacts the number of artifacts generated.
(02:17) Comparing Fragmentation Methods: Using a single DNA sample, the video compares three library preparation workflows:
Mechanical Fragmentation: Considered the gold standard, this method is random and unbiased, producing the lowest number of soft clipped reads.
Transposon-based Enzymatic: This fast method is prone to bias, often cutting near inverted repeats and creating significantly more soft clipped reads.
Endonuclease-based Enzymatic: A high-quality enzymatic option that produces far fewer artifacts than transposon methods, with performance much closer to mechanical fragmentation.
(04:19) Mechanism of Artifact Formation: The video explains a key hypothesis for how enzymatic methods create artifacts. Naturally occurring inverted repeats in the DNA can form stem-loop structures during library prep. These fragile structures can break and be incorrectly repaired, combining sequences from both original DNA strands into one, which results in a partially aligned (soft clipped) read.
(05:10) Identifying Artifacts with Bioinformatics: Free bioinformatics tools, such as FADE (Fragmentation, Artifact Detection, and Elimination), have been developed to specifically identify and filter out these artifactual soft clipped reads, preventing them from being misinterpreted as real variants.
(06:15) Impact and High-Risk Scenarios: While fewer than 1% of artifactual reads are mistakenly reported as variants, they can have a greater impact on sensitive applications like rare variant discovery. Target enrichment workflows are also more susceptible because they can concentrate any artifacts present in the targeted regions.
(06:41) Key Takeaways and Mitigation Strategies: To minimize the impact of soft clipped reads, it is crucial to:
Carefully select a library prep workflow, weighing the trade-offs between speed, DNA input, and potential for artifact generation.
Perform pilot tests on representative samples before embarking on large-scale projects.
Utilize appropriate analysis software to reliably identify and remove artifacts, ensuring the accuracy of variant calls.