Background Blood-based methods using cell-free DNA (cfDNA) are under development as an alternative to existing screening tests. However, early-stage detection of cancer using tumor-derived cfDNA has proven challenging because of the small proportion of cfDNA derived from tumor tissue in early-stage disease. A machine learning approach to discover signatures in cfDNA, potentially reflective of both tumor and non-tumor contributions, may represent a promising direction for the early detection of cancer. Methods Whole-genome sequencing was performed on cfDNA extracted from plasma samples ( N = 546 colorectal cancer and 271 non-cancer controls). Reads aligning to protein-coding gene bodies were extracted, and read counts were normalized. cfDNA tumor fraction was estimated using IchorCNA. Machine learning models were trained using k-fold cross-validation and confounder-based cross-validations to assess generalization performance. Results In a colorectal cancer cohort heavily weighted towards early-stage cancer (80% stage I/II), we achieved a mean AUC of 0.92 (95% CI 0.91–0.93) with a mean sensitivity of 85% (95% CI 83–86%) at 85% specificity. Sensitivity generally increased with tumor stage and increasing tumor fraction. Stratification by age, sequencing batch, and institution demonstrated the impact of these confounders and provided a more accurate assessment of generalization performance. Conclusions A machine learning approach using cfDNA achieved high sensitivity and specificity in a large, predominantly early-stage, colorectal cancer cohort. The possibility of systematic technical and institution-specific biases warrants similar confounder analyses in other studies. Prospective validation of this machine learning method and evaluation of a multi-analyte approach are underway. Electronic supplementary material The online version of this article (10.1186/s12885-019-6003-8) contains supplementary material, which is available to authorized users.

Neuronal ceroid lipofuscinosis type 2 (CLN2 disease) is an autosomal recessive condition caused by variants in the TPP1 gene, leading to deficient activity of the lysosomal enzyme tripeptidyl peptidase I (TPP1). We update on the spectrum of TPP1 variants associated with CLN2 disease, comprising 131 unique variants from 389 individuals (717 alleles) collected from the literature review, public databases, and laboratory communications. Previously unrecorded individuals were added to the UCL TPP1‐specific database. Two known pathogenic variants, c.509–1 G>C and c.622 C>T (p.(Arg208*)), collectively occur in 60% of affected individuals in the sample, and account for 50% of disease‐associated alleles. At least 86 variants (66%) are private to single families. Homozygosity occurs in 45% of individuals where both alleles are known (87% of reported individuals). Atypical CLN2 disease, TPP1 enzyme deficiency with disease onset and/or progression distinct from classic late‐infantile CLN2, represents 13% of individuals recorded with associated phenotype. NCBI ClinVar currently holds records for 37% of variants collected here. Effective CLN2 disease management requires early diagnosis; however, irreversible neurodegeneration occurs before a diagnosis is typically reached at age 5. Timely classification and public reporting of TPP1 variants is essential as molecular testing increases in use as a first‐line diagnostic test for pediatric‐onset neurological disease.

Maroteaux–Lamy syndrome (MPS VI) is an autosomal recessive lysosomal storage disorder caused by pathogenic ARSB gene variants, commonly diagnosed through clinical findings and deficiency of the arylsulfatase B (ASB) enzyme. Detection of ARSB pathogenic variants can independently confirm diagnosis and render genetic counseling possible. In this review, we collect and summarize 908 alleles (201 distinct variants, including 3 polymorphisms previously considered as disease‐causing variants) from 478 individuals diagnosed with MPS VI, identified from literature and public databases. Each variant is further analyzed for clinical classification according to American College of Medical Genetics and Genomics (ACMG) guidelines. Results highlight the heterogeneity of ARSB alleles, with most unique variants (59.5%) identified as missense and 31.7% of unique alleles appearing once. Only 18% of distinct variants were previously recorded in public databases with supporting evidence and clinical significance. ACMG recommends publishing clinical and biochemical data that accurately characterize pathogenicity of new variants in association with reporting specific alleles. Variants analyzed were sent to ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/), and MPS VI locus‐specific database (http://mps6-database.org) where they will be available. High clinical suspicion coupled with diagnostic testing for deficient ASB activity and timely submission and classification of ARSB variants with biochemical and clinical data in public databases is essential for timely diagnosis of MPS VI.

Background Duchenne muscular dystrophy (Duchenne) is caused by pathogenic variants in the DMD gene. Antisense oligonucleotides (AONs) are one emerging precision medicine treatment for Duchenne. DMD molecular genetic testing results guide precision‐therapy molecular eligibility, requiring healthcare providers to perform analyses currently uncommon in clinical laboratory and medical practices. Clear DMD variant notation and interpretation are key components of clinical care with the availability of precision medicine. Methods The DMD Open‐access Variant Explorer (DOVE) is a web‐based aid for DMD variant interpretation which additionally reports variant‐specific predicted molecular eligibility for therapy. DOVE was developed in Python and adapted to the Django Web framework, integrates existing open‐access tools, and does not rely on previous variant report/classification. Results DOVE [http://www.dmd.nl/DOVE] interprets colloquial and HGMD inputs of DMD variants to output HGMD variant nomenclature, theoretical molecular eligibility for therapy, and any predicted deleterious molecular consequences of therapy. DOVE relies on holistic in silico prediction of molecular eligibility for therapy in lieu of reference to an empirically defined, “variant‐eligible” list. Examples illustrate the advantage and necessity for holistic variant interpretation. Conclusion DOVE may prove useful for variant interpretation both at patient‐level and in large‐scale programs such as newborn screening and has broad application in concept to molecular genetic test result interpretation.

No abstract