REDBot: Natural language process methods for clinical copy number variation reporting in prenatal and products of conception diagnosis

Whole exome sequencing is an alternative method in the diagnosis of mitochondrial DNA diseases

BACKGROUND

Mitochondrial disease (MD) is genetically a heterogeneous group of disorders with impairment in respiratory chain complexes or pathways associated with the mitochondrial function. Nowadays, it is still a challenge for the genetic screening of MD due to heteroplasmy of mitochondrial genome and the complex model of inheritance. This study was designed to investigate the feasibility of whole exome sequencing (WES)-based testing as an alternative option for the diagnosis of MD.

METHODS

A Chinese Han cohort of 48 patients with suspect MD features was tested using nanoWES, which was a self-designed WES technique that covered the complete mtDNA genome and 21,019 nuclear genes. Fourteen patients were identified with a single genetic variant and three with single deletion in mtDNA.

RESULTS

The heteroplasmy levels of variants in mitochondrial genome range from 11% to 100%. NanoWES failed to identify multiple deletions in mtDNA compared with long range PCR and massively parallel sequencing (LR-PCR/MPS). However, our testing showed obvious advantages in identifying variations in nuclear DNA. Based on nanoWES, we identified two patients with nuclear DNA variation. One of them showed Xp22.33-q28 duplication, which indicated a possibility of Klinefelter syndrome.

CONCLUSION

NanoWES yielded a diagnostic rate of 35.4% for MD. With the rapid advances of next generation sequencing technique and decrease in cost, we recommend the usage of nanoWES as a first-line method in clinical diagnosis.

Cruxome: a powerful tool for annotating, interpreting and reporting genetic variants

Background

Next-generation sequencing (NGS) is an efficient tool used for identifying pathogenic variants that cause Mendelian disorders. However, the lack of bioinformatics training of researchers makes the interpretation of identified variants a challenge in terms of precision and efficiency. In addition, the non-standardized phenotypic description of human diseases also makes it difficult to establish an integrated analysis pathway for variant annotation and interpretation. Solutions to these bottlenecks are urgently needed.

Results

We develop a tool named “Cruxome” to automatically annotate and interpret single nucleotide variants (SNVs) and small insertions and deletions (InDels). Our approach greatly simplifies the current burdensome task of clinical geneticists and scientists to identify the causative pathogenic variants and build personal knowledge reference bases. The integrated architecture of Cruxome offers key advantages such as an interactive and user-friendly interface and the assimilation of electronic health records of the patient. By combining a natural language processing algorithm, Cruxome can efficiently process the clinical description of diseases to HPO standardized vocabularies. By using machine learning, in silico predictive algorithms, integrated multiple databases and supplementary tools, Cruxome can automatically process SNVs and InDels variants (trio-family or proband-only cases) and clinical diagnosis records, then annotate, score, identify and interpret pathogenic variants to finally generate a standardized clinical report following American College of Medical Genetics and Genomics/ Association for Molecular Pathology (ACMG/AMP) guidelines. Cruxome also provides supplementary tools to examine and visualize the genes or variations in historical cases, which can help to better understand the genetic basis of the disease.

Conclusions

Cruxome is an efficient tool for annotation and interpretation of variations and dramatically reduces the workload for clinical geneticists and researchers to interpret NGS results, simplifying their decision-making processes. We present an online version of Cruxome, which is freely available to academics and clinical researchers. The site is accessible at http://114.251.61.49:10024/cruxome/.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-021-07728-6.

REDBot: Natural language process methods for clinical copy number variation reporting in prenatal and products of conception diagnosis

Background

Current copy number variation (CNV) identification methods have rapidly become mature. However, the postdetection processes such as variant interpretation or reporting are inefficient. To overcome this situation, we developed REDBot as an automated software package for accurate and direct generation of clinical diagnostic reports for prenatal and products of conception (POC) samples.

Methods

We applied natural language process (NLP) methods for analyzing 30,235 in‐house historical clinical reports through active learning, and then, developed clinical knowledge bases, evidence‐based interpretation methods and reporting criteria to support the whole postdetection pipeline.

Results

Of the 30,235 reports, we obtained 37,175 CNV‐paragraph pairs. For these pairs, the active learning approaches achieved a 0.9466 average F1‐score in sentence classification. The overall accuracy for variant classification was 95.7%, 95.2%, and 100.0% in retrospective, prospective, and clinical utility experiments, respectively.

Conclusion

By integrating NLP methods in CNVs postdetection pipeline, REDBot is a robust and rapid tool with clinical utility for prenatal and POC diagnosis.