Skip to content
Bioinformatics Daily
by Syed Lokman

21. NGS data quality control

1. Introduction

Next-Generation Sequencing (NGS) has revolutionized genomics and molecular biology, enabling high-throughput, cost-effective sequencing of DNA and RNA. As a bioinformatics student, understanding the intricacies of NGS data quality control is crucial for ensuring the reliability and accuracy of your analyses. This comprehensive guide will walk you through the essential concepts, tools, and practices in NGS data quality control, preparing you for real-world applications in research and industry.

2. Understanding NGS Data

Before diving into quality control, it’s essential to understand the nature of NGS data:

  • Raw Data Format: NGS machines typically output data in FASTQ format, which contains sequence reads and their corresponding quality scores.
  • Sequencing Platforms: Different platforms (e.g., Illumina, Ion Torrent, PacBio) have unique characteristics and error profiles.
  • Read Types: Single-end vs. paired-end reads, short reads vs. long reads.
  • Sequencing Depth: The number of times a nucleotide is read during the sequencing process.

3. Importance of Quality Control in NGS

Quality control is a critical step in the NGS data analysis pipeline for several reasons:

  • Error Detection: Identifying and filtering out low-quality reads and sequencing artifacts.
  • Bias Reduction: Minimizing systematic biases introduced during library preparation or sequencing.
  • Downstream Analysis Improvement: Ensuring high-quality input data for assembly, alignment, and variant calling.
  • Resource Optimization: Reducing computational resources wasted on processing poor-quality data.
  • Reproducibility: Establishing standardized quality thresholds for consistent and comparable results.

4. Key Quality Control Metrics

To effectively assess NGS data quality, you need to understand and interpret various metrics:

  1. Base Quality Scores (Phred Scores)

    • Definition: Logarithmic measure of base calling accuracy.
    • Interpretation: Q30 means 99.9% base call accuracy; higher is better.

  2. Per-base Sequence Quality

    • Assessment of quality scores across all bases at each position.
    • Helps identify position-specific quality drops.

  3. Per-sequence Quality Scores

    • Distribution of average quality scores for all reads.
    • Identifies subsets of low-quality reads.

  4. Sequence Length Distribution

    • Variation in read lengths.
    • Important for platforms with variable read lengths (e.g., PacBio, Nanopore).

  5. GC Content

    • Distribution of GC percentage across all reads.
    • Helps identify contamination or biased library preparation.

  6. Sequence Duplication Levels

    • Degree of read duplication.
    • High duplication may indicate PCR bias or low library complexity.

  7. Overrepresented Sequences

    • Frequently occurring sequences (e.g., adapters, contaminants).
    • Helps identify and remove technical artifacts.

  8. K-mer Content

    • Frequency of short subsequences.
    • Useful for detecting biases and contaminations.

5. Quality Control Tools and Software

Familiarize yourself with these essential tools for NGS data quality control:

  1. FastQC

    • De facto standard for quick quality checks.
    • Provides visual reports on various quality metrics.

  2. MultiQC

    • Aggregates results from multiple bioinformatics tools.
    • Useful for comparing samples and runs.

  3. Trimmomatic

    • Flexible read trimming tool for Illumina data.
    • Removes adapters and low-quality bases.

  4. Cutadapt

    • Removes adapter sequences, primers, and other unwanted sequences.
    • Supports various types of adapter trimming.

  5. BBTools Suite

    • Collection of fast, multithreaded bioinformatics tools.
    • Includes tools for read quality filtering, error correction, and more.

  6. PRINSEQ

    • Generates summary statistics and allows filtering, reformatting, and trimming.
    • Useful for both pre- and post-processing of sequence data.

  7. Kraken

    • Rapid taxonomic classification of sequence reads.
    • Helps identify contamination from other organisms.

6. Quality Control Workflow

A typical NGS data quality control workflow includes the following steps:

  1. Initial Quality Assessment

    • Run FastQC on raw data to identify quality issues.

  2. Adapter Trimming

    • Remove adapter sequences using tools like Cutadapt or Trimmomatic.

  3. Quality Trimming

    • Trim or filter out low-quality bases and reads.

  4. Contamination Screening

    • Use tools like Kraken to identify and remove contaminant sequences.

  5. Error Correction

    • Apply error correction algorithms, especially for long-read data.

  6. Post-processing Quality Check

    • Re-run FastQC to verify improvement in data quality.

  7. MultiQC Report Generation

    • Aggregate results for easy comparison across samples.

7. Use Cases

Understanding real-world applications will help you appreciate the importance of quality control:

7.1 Whole Genome Sequencing (WGS)

  • Scenario: You’re working on a human WGS project to identify genetic variants associated with a rare disease.
  • QC Focus:
    • High coverage uniformity
    • Low duplication rates
    • Minimal adapter contamination
  • Impact: Proper QC ensures accurate variant calling, reducing false positives that could mislead clinical interpretations.

7.2 RNA-Seq

  • Scenario: Analyzing differential gene expression in cancer samples.
  • QC Focus:
    • rRNA contamination
    • GC bias
    • 3’ bias in gene coverage
  • Impact: Good QC practices lead to more reliable quantification of gene expression levels and identification of truly differentially expressed genes.

7.3 Metagenomics

  • Scenario: Studying microbial community composition in soil samples.
  • QC Focus:
    • Removal of host DNA contamination
    • Identification of potential lab contaminants
    • Careful handling of low-abundance species
  • Impact: Proper QC is crucial for accurately representing the true microbial diversity and avoiding artifacts that could skew community analysis.

7.4 ChIP-Seq

  • Scenario: Mapping transcription factor binding sites across the genome.
  • QC Focus:
    • PCR duplication rates
    • Signal-to-noise ratio
    • Library complexity
  • Impact: Effective QC helps distinguish true binding events from background noise, leading to more accurate peak calling and motif discovery.

8. Advanced Topics

As you progress in your bioinformatics journey, consider exploring these advanced quality control topics:

  1. Machine Learning in QC

    • Using ML algorithms to predict and correct sequencing errors.
    • Automated classification of high-quality vs. low-quality samples.

  2. Long-Read Sequencing QC

    • Specific challenges and tools for PacBio and Oxford Nanopore data.
    • Error correction strategies for high error rates.

  3. Single-Cell Sequencing QC

    • Unique considerations for cell barcodes and UMIs.
    • Identifying and handling cell doublets and empty droplets.

  4. Structural Variant Detection QC

    • Ensuring sufficient read depth and insert size distribution.
    • Validating complex structural rearrangements.

  5. Cloud-based QC Pipelines

    • Implementing scalable QC workflows on cloud platforms.
    • Leveraging containerization for reproducible QC processes.

  6. Quality Control in Clinical NGS

    • Stricter QC requirements for diagnostic applications.
    • Compliance with regulatory standards (e.g., CLIA, CAP).

9. Conclusion

Mastering NGS data quality control is fundamental for any aspiring bioinformatician. By understanding the principles, metrics, tools, and workflows discussed in this guide, you’ll be well-equipped to ensure the reliability and reproducibility of your genomic analyses. Remember that quality control is not just a box to check—it’s an ongoing process that requires critical thinking and adaptation to each unique project. As sequencing technologies continue to evolve, stay curious and keep updating your QC knowledge and skills.

To further your learning, consider:

  • Practicing with public datasets from repositories like SRA or ENA.
  • Contributing to open-source QC tools or developing your own.
  • Staying updated with the latest literature on NGS quality control methods.

By investing time in mastering NGS data quality control, you’re laying a solid foundation for a successful career in bioinformatics and genomics research.