Assessing concordance among human, in silico predictions and functional assays on genetic variant classification.

MOTIVATION: A variety of in silico tools have been developed and frequently used to aid high-throughput rapid variant classification, but their performances vary, and their ability to classify variants of uncertain significance were not systemically assessed previously due to lack of validation data. This has been changed recently by advances of functional assays, where functional impact of genetic changes can be measured in single-nucleotide resolution using saturation genome editing (SGE) assay. RESULTS: We demonstrated the neural network model AIVAR (Artificial Intelligent VARiant classifier) was highly comparable to human experts on multiple verified datasets. Although highly accurate on known variants, AIVAR together with CADD and PhyloP showed non-significant concordance with SGE function scores. Moreover, our results indicated that neural network model trained from functional assay data may not produce accurate prediction on known variants. AVAILABILITY AND IMPLEMENTATION: All source code of AIVAR is deposited and freely available at https://github.com/TopGene/AIvar. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Assessing the Variations in Breast/Ovarian Cancer Risk for Chinese BRCA1/2 Carriers.

Background: Cancer risks vary in different BRCA1/2 mutations. We are interested in identifying regions associated with elevated/reduced risks of breast/ovarian cancers in the Chinese population and comparing with previously reported Caucasian-based breast/ovarian cancer cluster regions (OCCR/BCCR). We also aim to characterize the distribution and estimate the cancer risks of different Chinese recurrent mutations. Methods: A total of 3,641 cancer-free women and 4,278 female cancer patients were included in the study. Germline BRCA1/2 status was detected with amplicon-based next-generation sequencing. We calculated the odds ratio (OR) of breast cancer and OR of ovarian cancer, and their ratio of the two ORs (ROR) for each region. ROR >1 indicated elevated odds of breast cancer and/or decreasing odds of ovarian cancer, and vice versa. The frequency, distribution, and penetrance of six known Chinese founder mutations were characterized, respectively. Haplotype analysis and age estimation were performed on the most prevalent founder mutation BRCA1: c.5470_5477del. Results: A total of 729 subjects were detected with germline BRCA1/2 deleterious mutations. The putative Chinese OCCR/BCCR partially overlapped with Caucasian-based OCCR/BCCR and shared structural-functional characteristics. The six known Chinese founder mutations greatly vary in both distribution and penetrance. The two widely spread mutations are estimated to convey low penetrance, while the area-restricted founder mutations seemed to confer higher/complete penetrance. BRCA1: c.5470_5477del is estimated to have emerged ∼2,090 years ago (70 B.C.) during the Han dynasty. Conclusions: BRCA1/2 carriers with different genotypes have significantly different cancer risks. An optimal risk assessment should be mutation specific, rather than concerning a single figure.

Retrospective reinterpretation and reclassification of BRCA1/2 variants from Chinese population.

BACKGROUND: The accurate interpretation of BRCA1/2 variants becomes increasingly important in breast cancer and other related cancers including ovarian cancer, prostate cancer, pancreatic cancer and so forth. In the past decades, especially before year 2015, limitations of techniques and lack of databases and guidelines have led to possible misinterpretation of the clinical significance of sequence variants of BRCA1/2. A published study reported reclassification of some BRCA1/2 variants previously classified as variants of uncertain significance (VUS) to likely pathogenic in breast or ovarian cancer patients from Korea. However, little is known about the situation in Chinese population. METHODS: We retrospectively retrieved 109 publications studying about BRCA1/2 variants of Chinese population from the year 1999 to year 2019 (March). After excluding publications of meta-analysis and publications with missing data, 72 publications were eventually retained for subsequent analysis. In total, 1,351 BRCA variants (673 BRCA1 variants and 678 BRCA2 variants) derived from 42,430 Chinese cancer patients were standardized and reinterpreted using ACMG/AMP 2015 guidelines and China Expert Consensus on BRCA variant interpretation by genetic counselors. RESULTS: Among the 1,351 BRCA variants, the majority of interpretation (91.7%, 1,239/1,351) remained the same as previously published. However, there were 112 (8.3%, 112/1,351) variants (64 BRCA1, 48 BRCA2) reclassified with different categories. CONCLUSIONS: Our results demonstrated that clinical significance of not only VUS, but also pathogenic/likely pathogenic variants varied from time to time in the Chinese population. Precise reinterpretation of BRCA1/2 variants is of crucial importance to genetic counseling or clinical decision-making for risk individuals or patients.

Molecular Profiling Reveals Common and Specific Development Processes in Different Types of Gynecologic Cancers.

BACKGROUND: Gynecologic cancers have become a major threat to women's health. The molecular biology of gynecologic cancers is not as well understood as that of breast cancer, and precision targeting is still new. Although viewed collectively as a group of cancers within the female reproductive system, they are more often studied separately. A comprehensive within-group comparison on molecular profiles is lacking. METHODS: We conducted a whole-exome sequencing study of cervical/endometrial/ovarian cancer samples from 209 Chinese patients. We combined our data with genomic and transcriptomic data from relevant TCGA cohorts to identify and verify common/exclusive molecular changes in cervical/endometrial/ovarian cancer. RESULTS: We identified shared molecular features including a COSMIC signature of deficient mismatch repair (dMMR), four recurrent copy-number variation (CNV) events, and extensive alterations in PI3K-Akt-mTOR signaling and cilium component genes; we also identified transcription factors and pathways that are exclusively altered in cervical/endometrial/ovarian cancer. The functions of the commonly/exclusively altered genomic circuits suggest (1) a common reprogramming process during early tumor initiation, which involves PI3K activation, defects in mismatch repair and cilium organization, as well as disruption in interferon signaling and immune recognition; (2) a cell-type specific program at late-stage tumor development that eventually lead to tumor proliferation and migration. CONCLUSION: This study describes, from a molecular point of view, how similar and how different gynecologic cancers are, and it provides a hypothesis about the causes of the observed similarities and differences.