Library Complexity

The absolute number of distinct, non-duplicate DNA fragments in a library before amplification. This number sets the theoretical upper limit for variant detection sensitivity.

• Pre-PCR Diversity: Measured by counting unique molecular identifiers (UMIs) after deduplication.
• Saturation Point: The sequencing depth at which no new unique molecules are observed.
• Direct Correlation: A library with 10,000 unique genomes cannot reliably detect a variant at 0.001% frequency.

The fraction of sequenced reads that are identical copies of the same original molecule, generated during PCR amplification. High duplication indicates low input material or over-amplification.

• Computational Deduplication: UMIs are used to collapse reads sharing the same barcode and genomic coordinates into a single consensus read.
• Inverse Metric: Duplication rate is the inverse of library complexity; a 90% duplication rate means only 10% of reads are informative.
• Optical Duplicates: A separate artifact class caused by cluster misidentification on the sequencer flow cell, filtered without UMIs.

The range and frequency of DNA fragment lengths in a library, reflecting both the biological fragmentation process and the physical limits of library preparation chemistry.

• cfDNA Signature: Cell-free DNA naturally peaks at ~166 bp, corresponding to nucleosome-protected DNA.
• Complexity Proxy: A narrow, sharp peak suggests low complexity from a single source; a broad, smooth distribution indicates high diversity.
• PCR Bottleneck: Over-amplification skews the distribution toward shorter fragments due to preferential amplification of smaller templates.

The evenness of sequencing coverage across targeted genomic regions. Poor uniformity forces over-sequencing to achieve minimum depth at cold spots, wasting reads and reducing effective complexity.

• GC Bias: The primary driver of non-uniformity, where extreme guanine-cytosine content causes dropouts. Corrected via GC bias correction algorithms.
• Fold-80 Penalty: A standard metric quantifying the extra sequencing required to bring 80% of targets to the mean coverage depth.
• Probe Design: Poorly tiled capture probes create systematic gaps that no amount of sequencing can fill, permanently capping complexity.

The precise measurement of amplifiable DNA molecules entering library preparation, typically using digital droplet PCR (ddPCR) or fluorometric methods. This defines the absolute ceiling of possible complexity.

• Genome Equivalents: 3.3 pg of human DNA equals one haploid genome equivalent. A 10 ng input contains ~3,000 genome equivalents.
• Conversion Efficiency: The fraction of input molecules successfully ligated to adapters. Low efficiency directly destroys complexity.
• Critical for Liquid Biopsy: When only nanograms of cfDNA are available, every molecule lost during library prep permanently erases a potential biomarker.

The degree to which the relative abundance of original molecules is maintained through library preparation and sequencing. Distortion creates quantitative bias.

• PCR Jackpotting: Stochastic early-cycle amplification where a single molecule is disproportionately copied, creating a false clonal expansion.
• Adapter Ligation Bias: Sequence-dependent efficiency differences in adapter attachment that systematically deplete certain genomic regions.
• Quantitative Integrity: Essential for variant allele frequency (VAF) accuracy. A true 0.5% VAF must remain 0.5% after library construction, not 2% due to jackpotting.