Key Deliverables & Statistical Frameworks
Per-base Phred quality diagnostics, GC bias detection, library duplication, adapter read-through, insert size profiling, and long-read Nanopore QC.
[DOMAIN 8: SEQUENCING QUALITY CONTROL & PREPROCESSING]
Plot 098 — FastQC Per-Base Phred Quality Boxplot
Use Case Application: Diagnoses sequencing base-call accuracy decay toward 3' read ends to guide quality trimming parameters.
Required Input Data: Raw FASTQ sequencing files parsed across read cycle positions.
Statistical & Algorithmic Method: Per-cycle median, interquartile range (IQR), and 10th-90th percentile Phred scores Q = -10 * log10(P_error).
[DOMAIN 8: SEQUENCING QUALITY CONTROL & PREPROCESSING]
Plot 099 — Per-Sequence GC-Content Distribution Curve
Use Case Application: Detects cross-species microbial contamination, rRNA carryover, or PCR amplification bias in sequencing libraries.
Required Input Data: FASTQ reads evaluated for per-read guanine-cytosine (GC%) percentage.
Statistical & Algorithmic Method: Empirical GC% density curve compared against a fitted theoretical Normal distribution; multimodal peaks indicate contamination.
[DOMAIN 8: SEQUENCING QUALITY CONTROL & PREPROCESSING]
Plot 100 — Sequence Duplication Level Decay Curve
Use Case Application: Distinguishes low-complexity over-amplified PCR libraries from true high-depth transcriptomic coverage.
Required Input Data: FASTQ read sequences evaluated for exact or deduplicated copy multiplicity.
Statistical & Algorithmic Method: Proportion of library reads binned by duplication level (1x, 2x, ..., >10x) before and after UMI/optical deduplication.
[DOMAIN 8: SEQUENCING QUALITY CONTROL & PREPROCESSING]
Plot 101 — Adapter Read-Through Content Profile
Use Case Application: Identifies short insert fragments reading into 3' sequencing adapters that must be trimmed prior to alignment.
Required Input Data: FASTQ read cycles matched against Illumina/BGI/ONT adapter k-mer sequences.
Statistical & Algorithmic Method: Cumulative percentage of reads containing adapter k-mers at each base cycle position.
[DOMAIN 8: SEQUENCING QUALITY CONTROL & PREPROCESSING]
Plot 102 — Paired-End Fragment Insert Size Histogram
Use Case Application: Verifies library size-selection accuracy and detects structural deletion/insertion shifts in paired-end sequencing.
Required Input Data: Coordinate-sorted paired-end BAM alignments (Picard CollectInsertSizeMetrics).
Statistical & Algorithmic Method: Empirical distribution of properly paired read template lengths (TLEN) summarized by median and Median Absolute Deviation (MAD).
[DOMAIN 8: SEQUENCING QUALITY CONTROL & PREPROCESSING]
Plot 103 — Relative Log Expression (RLE) Normalization Boxplot
Use Case Application: Confirms removal of library-size and batch technical artifacts across samples after TMM, DESeq2, or RUV normalization.
Required Input Data: Log-transformed raw vs. normalized expression matrix across all samples.
Statistical & Algorithmic Method: Per-gene median subtraction RLE_ig = log2(y_ig) - median_j(log2(y_jg)); well-normalized samples center tightly around zero.
[DOMAIN 8: SEQUENCING QUALITY CONTROL & PREPROCESSING]
Plot 104 — Sample-to-Sample Replicate Correlation Heatmap
Use Case Application: Verifies biological replicate reproducibility and flags sample swaps or outliers prior to downstream modeling.
Required Input Data: Normalized gene expression, methylation, or proteomics matrix across all biological and technical replicates.
Statistical & Algorithmic Method: Pairwise Pearson or Spearman rank correlation matrix clustered hierarchically across samples.
[DOMAIN 8: SEQUENCING QUALITY CONTROL & PREPROCESSING]
Plot 105 — Oxford Nanopore Read Length vs. Quality 2D Density
Use Case Application: Evaluates long-read N50 continuity and base-call Q-score yield in Oxford Nanopore or PacBio HiFi runs.
Required Input Data: Sequencing summary table (sequencing_summary.txt) containing read sequence length and mean Q-score per read.
Statistical & Algorithmic Method: 2D bivariate histogram / kernel density of log10(read length) vs. mean Phred Q-score with read N50 threshold overlay.