Cell-free DNA screening for fetal aneuploidy — from maternal blood draw through massively parallel sequencing to chromosome-level Z-score risk calling
Non-invasive prenatal testing (NIPT) exploits a simple biological fact: throughout pregnancy, apoptotic trophoblast cells from the placenta continuously shed short fragments of DNA into the maternal bloodstream. These placental-derived cell-free DNA (cfDNA) fragments — genetically representative of the fetus in the overwhelming majority of pregnancies — mix with the mother's own cfDNA, and can be captured, sequenced, and analyzed from a single tube of maternal blood as early as 10 weeks gestation.
Circulating cfDNA in maternal plasma is not fetal in the strict sense — it originates from apoptosis of cytotrophoblast cells at the maternal-placental interface, released continuously as the placenta remodels throughout gestation.
Key biological features: • Fragment length: cfDNA is cleaved during apoptosis by endonucleases acting on nucleosome-packaged chromatin, producing a characteristic ~140–160bp fragment size (shorter than maternal-origin fragments on average) • Half-life: placental cfDNA clears from maternal circulation within minutes to hours of delivery, confirming it is a real-time signal of the current pregnancy, not a persistent maternal store • Detectable from ~9–10 weeks: sufficient placental mass and turnover must accumulate before fetal fraction is reliably measurable • Represents placenta, not fetus directly: in the rare case of confined placental mosaicism, cfDNA signal reflects placental genotype, which may differ from the true fetal karyotype
Sample handling requirements: • Blood collected into specialized cfDNA-stabilizing tubes (contain a fixative that prevents maternal leukocyte lysis, which would otherwise flood the sample with maternal genomic DNA and dilute fetal fraction) • Plasma separated by double centrifugation, typically within 24–72 hours if using stabilized tubes • Extracted total cfDNA pool is a mixture: ~85–90% maternal-origin, ~10–15% placental-origin in a typical singleton pregnancy
Every NIPT result is only as reliable as the fraction of placental DNA available to analyze. Fetal fraction (FF) must be estimated before — or alongside — the aneuploidy analysis itself, because a low FF fundamentally limits the statistical power to detect an abnormal chromosome dosage signal.
How fetal fraction is estimated: • Fragment-size-based methods: placental fragments are on average slightly shorter than maternal fragments, allowing size-distribution deconvolution • SNP-based methods: compare maternal-only genotype (from buffy coat) against the mixed plasma signal at informative SNP loci • Y-chromosome-based (male fetuses only): read depth on chrY directly estimates FF, but is not generalizable to female fetuses
Consequences of low fetal fraction: • Below the assay-specific reportable threshold (commonly ~4%), the aneuploidy signal cannot be reliably separated from noise → result reported as "no call" / test failure rather than a false reassurance • No-call rate in general obstetric populations: roughly 1–5%, higher in the causes listed below • Redraw improves fetal fraction in most cases as gestational age advances, but a subset of patients will have persistently low FF
Common causes of low fetal fraction: • High maternal body mass index / weight — thought to reflect dilution by increased maternal blood volume and greater maternal cfDNA turnover from adipose tissue • Early gestational age — placental mass and turnover have not yet reached sufficient levels • Certain aneuploidies (notably trisomy 18 and trisomy 13) are statistically associated with smaller placentas and lower average fetal fraction than euploid or trisomy 21 pregnancies • Sample handling delays, assay/technical factors
A persistently low fetal fraction is itself a soft risk marker — some studies show a modestly increased risk of aneuploidy and adverse pregnancy outcomes in pregnancies with unexplained very low FF, independent of the aneuploidy call.
Once extracted, the mixed maternal/placental cfDNA pool is prepared into a sequencing library and read on a massively parallel (next-generation) sequencing platform. Rather than trying to distinguish individual fetal reads from maternal reads directly, most clinical NIPT assays rely on a population-level counting strategy: millions of short reads are aligned to the reference genome, tallied by chromosome, and compared statistically to expected proportions.
Three broad analytic strategies are used commercially, all converging on the same goal: quantify chromosome dosage in the mixed cfDNA pool with enough precision to detect a fractional excess or deficit contributed by an aneuploid fetal chromosome.
1. Whole-genome (shotgun) massively parallel sequencing (MPS): • Every fragment in the library is sequenced without target enrichment • Reads aligned to reference genome (GRCh37/38), uniquely mapping reads retained • Reads binned into genomic windows and summed per chromosome • Advantage: unbiased genome-wide coverage, extensible to genome-wide microdeletion screening • Disadvantage: majority of sequencing effort spent on chromosomes never being tested
2. Targeted (chromosome-selective) sequencing: • PCR-amplifies specific loci on chromosomes of interest (13, 18, 21, X, Y) before sequencing • More efficient use of sequencing depth, lower cost per sample • Disadvantage: cannot be repurposed for genome-wide analysis
3. SNP-based sequencing: • Targets thousands of single-nucleotide polymorphism loci • Directly models the ratio of maternal and fetal alleles at each SNP rather than only total chromosome dosage • Enables detection of triploidy and vanishing twin/multiple gestation confounders that dosage-only methods can miss • Also used for fetal fraction estimation without needing a male fetus
Read alignment and quality control: • GC-content correction: sequencing efficiency varies by local GC content; raw counts are normalized against a GC-content curve built from reference samples • Duplicate read removal, mapping quality filtering • Each chromosome's corrected read count is expressed as a fraction of total mapped reads — the "chromosomal representation"
A raw chromosome read count means nothing on its own — it must be compared against the distribution of values seen in a large reference set of confirmed-euploid pregnancies. The Z-score expresses how many standard deviations a given sample's chromosomal representation lies from that euploid reference mean, and it is the statistical backbone of virtually every commercial NIPT algorithm.
For a target chromosome (say, chromosome 21):
Z = (observed chromosomal representation − reference population mean) / reference population standard deviation
• The reference mean and standard deviation are established from thousands of sequencing runs on samples with confirmed euploid outcomes, run on the same platform and normalized the same way • An extra copy of chromosome 21 (trisomy 21) contributes proportionally more chr21-derived fragments to the pool → observed representation shifts upward → Z-score becomes strongly positive • The magnitude of the shift is directly proportional to fetal fraction: a fetus with trisomy 21 at 15% fetal fraction produces a much larger, more separable Z-score shift than the same trisomy at 4% fetal fraction — this is why fetal fraction and Z-score are analyzed jointly, not independently
Why chromosomes 21, 18, and 13 are the primary targets: • These are the only autosomal trisomies commonly compatible with live birth, making them the highest-yield targets for a screening test intended to inform ongoing pregnancy management • Chromosome 21 is gene-poor and GC-content-favorable for sequencing normalization, contributing to its high assay performance • Sex chromosome analysis (X, Y) additionally screens for monosomy X (Turner syndrome), XXY (Klinefelter), XYY, and XXX, though performance characteristics are less robust than autosomal trisomy detection
Sources of Z-score noise and false signals: • Confined placental mosaicism — a chromosomal error present only in the placenta, not the fetus, still shifts cfDNA representation • Vanishing twin — a co-twin demise contributes residual cfDNA that can persist for weeks and confound the surviving twin's signal • Maternal mosaicism or an undiagnosed maternal malignancy — maternal cfDNA itself can carry an abnormal copy-number signal unrelated to the fetus • Maternal copy-number variants overlapping the tested chromosome region
A Z-score is a statistical distance, not a diagnosis. A |Z|>3 result on chromosome 21 substantially raises the probability of trisomy 21 but does not confirm it — the test remains a screen, and every high-risk NIPT result requires diagnostic confirmation before any irreversible clinical decision is made.
The final NIPT report translates a statistical Z-score into a categorical result — low-risk or high-risk — but responsible reporting must also communicate what that result actually means for a given patient. Because Down syndrome and other aneuploidies are relatively rare in the general population, even a highly sensitive and specific test carries a positive predictive value that varies enormously with the patient's prior (a priori) risk, most commonly driven by maternal age.
Sensitivity and specificity are properties of the test itself; positive predictive value (PPV) also depends on how common the condition is in the population being tested (Bayes' theorem in action).
Worked intuition: • In a low a priori risk patient (young, no other risk factors), the pre-test probability of trisomy 21 might be roughly 1 in 1,000. Even with 99% sensitivity and 99.9% specificity, the small pool of true positives is diluted by the far larger pool of unaffected pregnancies that still generate a small number of false positives — pushing PPV to roughly 30–50% • In a high a priori risk patient (advanced maternal age, prior abnormal screen, ultrasound soft markers), the pre-test probability might be 1 in 100 or higher — the same test now yields a PPV well above 80–90%, because the true-positive pool is proportionally much larger relative to the fixed false-positive rate
Practical implication: a "high-risk" NIPT result in a 22-year-old with no other risk factors is meaningfully less likely to represent a true positive than the identical "high-risk" result in a 42-year-old — the laboratory report should ideally state a patient-specific PPV, not just the categorical call.
Causes of false-positive and false-negative results: • Confined placental mosaicism (placenta abnormal, fetus normal) — most common explanation for a false-positive • Vanishing twin with an aneuploid co-twin • Maternal mosaicism, including undiagnosed maternal sex-chromosome mosaicism • Maternal malignancy — rare but documented cause of an anomalous genome-wide cfDNA signal, occasionally the first clue leading to maternal cancer diagnosis • True fetal mosaicism • Laboratory/technical error, low fetal fraction pushing a borderline true result across threshold noise
Because NIPT is a screening test, professional guidelines (ACOG, ISPD, ACMG) uniformly recommend that no irreversible pregnancy management decision be made on the basis of a high-risk NIPT result alone.
Confirmatory pathway: • Chorionic villus sampling (first trimester) or amniocentesis (second trimester) with karyotype and/or chromosomal microarray remains the diagnostic standard • Genetic counseling should accompany both the pretest discussion and any high-risk result, covering the screening-versus-diagnostic distinction, possible causes of a false-positive, and reproductive options
Expanded panels and microdeletion screening: • Many commercial platforms now offer add-on screening for microdeletion syndromes (e.g., 22q11.2 deletion / DiGeorge syndrome, 1p36 deletion, Cri-du-chat) • Because these conditions are individually far rarer than trisomy 21, the same test performance characteristics yield a much lower PPV — a substantial proportion of "high-risk" microdeletion calls do not represent a true affected fetus • Sex chromosome aneuploidy screening (Turner, Klinefelter, XYY, XXX) has more heterogeneous performance than the core autosomal panel and a wider range of phenotypic severity, which should be discussed pretest • Twin pregnancies, egg-donor pregnancies, and pregnancies following recent transfusion or stem-cell transplant all carry assay-specific caveats that should be reviewed before ordering
The consistent message across all major guidelines: NIPT is the most effective aneuploidy screening test available today, but it is still a screening test. A low-risk result substantially reassures; a high-risk result substantially raises concern — but only a diagnostic test (karyotype or microarray obtained via CVS or amniocentesis) confirms a chromosomal diagnosis.