Recently, convolutional neural networks (CNNs) have been widely applied in the seismic facies classification. However, even state-of-the-art CNN architectures often encounter classification confusion ...distinguishing seismic facies at their boundaries. In addition, the annotation is a highly time-consuming task, especially when dealing with 3-D seismic data volumes. While traditional semisupervised methods reduce dependence on annotation, they are susceptible to interference from unreliable pseudolabels. To address these challenges, we propose a semisupervised seismic facies classification method called CONSS, which effectively mitigates classification confusion through contrastive learning. Our proposed method requires only 1% of labeled data, significantly reducing the demand for annotation. To minimize the influence of unreliable pseudolabels, we also introduce a confidence strategy to select positive and negative sample pairs from reliable regions for contrastive learning. Experimental results on the publicly available seismic datasets, the Netherlands F3 and SEAM AI challenge datasets, demonstrate that the proposed method outperforms classic semisupervised methods, including self-training and consistency regularization, achieving exceptional classification performance.
Abstract
Current non-invasive prenatal screening (NIPS) analyzes circulating fetal cell-free DNA (cfDNA) in maternal peripheral blood for selected aneuploidies or microdeletion/duplication syndromes. ...Many genetic disorders are refractory to NIPS largely because the maternal genetic material constitutes most of the total cfDNA present in the maternal plasma, which hinders the detection of fetus-specific genetic variants. Here, we developed an innovative sequencing method, termed
co
ordinative
a
llele-aware
t
arget
e
nrichment
seq
uencing (COATE-seq), followed by multidimensional genomic analyses of sequencing read depth, allelic fraction, and linked single nucleotide polymorphisms, to accurately separate the fetal genome from the maternal background. Analytical confounders including multiple gestations, maternal copy number variations, and absence of heterozygosity were successfully recognized and precluded for fetal variant analyses. In addition, fetus-specific genomic characteristics, including the cfDNA fragment length, meiotic error origins, meiotic recombination, and recombination breakpoints were identified which reinforced the fetal variant assessment. In 1129 qualified pregnancies tested, 54 fetal aneuploidies, 8 microdeletions/microduplications, and 8 monogenic variants were detected with 100% sensitivity and 99.3% specificity. Using the comprehensive cfDNA genomic analysis tools developed, we found that 60.3% of aneuploidy samples had aberrant meiotic recombination providing important insights into the mechanism underlying meiotic nondisjunctions. Altogether, we show that the genetic deconvolution of the fetal and maternal cfDNA enables thorough and accurate delineation of fetal genome which paves the way for the next-generation prenatal screening of essentially all types of human genetic disorders.
Recently, convolutional neural networks (CNNs) have been widely applied in the seismic facies classification. However, even state-of-the-art CNN architectures often encounter classification confusion ...distinguishing seismic facies at their boundaries. Additionally, the annotation is a highly time-consuming task, especially when dealing with 3D seismic data volumes. While traditional semi-supervised methods reduce dependence on annotation, they are susceptible to interference from unreliable pseudo-labels. To address these challenges, we propose a semi-supervised seismic facies classification method called CONSS, which effectively mitigates classification confusion through contrastive learning. Our proposed method requires only 1% of labeled data, significantly reducing the demand for annotation. To minimize the influence of unreliable pseudo-labels, we also introduce a confidence strategy to select positive and negative sample pairs from reliable regions for contrastive learning. Experimental results on the publicly available seismic datasets, Netherlands F3 and SEAM AI challenge dataset, demonstrate that the proposed method outperforms classic semi-supervised methods, including self-training and consistency regularization, achieving exceptional classification performance. Our all codes and data are available at https://github.com/upcliuwenlong/CONSS_SEISMIC_FACIES .
Multigene panel testing of breast cancer predisposition genes have been extensively conducted in Europe and America, which is relatively rare in Asia however. In this study, we assessed the frequency ...of germline mutations in 40 cancer predisposition genes, including BRCA1 and BRCA2, among a large cohort of Chinese patients with high hereditary risk of BC. From 2015 to 2016, consecutive BC patients from 26 centers of China with high hereditary risk were recruited (n = 937). Clinical information was collected and next‐generation sequencing (NGS) was performed using blood samples of participants to identify germline mutations. In total, we acquired 223 patients with putative germline mutations, including 159 in BRCA1/2, 61 in 15 other BC susceptibility genes and 3 in both BRCA1/2 and non‐BRCA1/2 gene. Major mutant non‐BRCA1/2 genes were TP53 (n = 18), PALB2 (n = 11), CHEK2 (n = 6), ATM (n = 6) and BARD1 (n = 5). No factors predicted pathologic mutations in non‐BRCA1/2 genes when treated as a whole. TP53 mutations were associated with HER‐2 positive BC and younger age at diagnosis; and CHEK2 and PALB2 mutations were enriched in patients with luminal BC. Among high hereditary risk Chinese BC patients, 23.8% contained germline mutations, including 6.8% in non‐BRCA1/2 genes. TP53 and PALB2 had a relatively high mutation rate (1.9 and 1.2%). Although no factors predicted for detrimental mutations in non‐BRCA1/2 genes, some clinical features were associated with mutations of several particular genes.
What's new?
The prevalence of mutations in breast cancer predisposition genesare not well investigated in Asia. We assessed germline mutations of 40 cancer susceptibility genes in 937 consecutive selected breast cancer patients from 26 centers of China, and discovered 23.8% of participates carried the pathogenic mutation, including 6.8% with mutations in non‐BRCA1/2 genes, while TP53 and PALB2 had a relatively high mutation rates (1.9% and 1.2%).There was no factors predicted for detrimental mutations in non‐BRCA1/2 genes when treated as a whole.
Along with the dramatic development of molecular diagnostic testing for the detection of oncogene variations, reference materials (RMs) have become increasingly important in performance evaluation of ...genetic testing.
In this study, we built a set of RMs for genetic testing based on next-generation sequencing (NGS).
Solid tumor tissues were selected as the samples of RMs for preparation. NGS was used to determine and validate the variants and the mutation frequency in DNA samples. Digital PCR was used to determine the copy numbers of RNA samples. The performance of the RMs was validated by six laboratories.
Thirty common genetic alterations were designed based on these RMs. RMs consisted of a positive reference, a limit of detection reference, and a negative reference. The validation results confirmed the performance of the RMs.
These RMs may be an attractive tool for the development, validation, and quality monitoring of molecular genetic testing.
The absence of interpretation guidelines and limited data on BRCA1/2 mutations in the Chinese population have impeded the detection of BRCA variants based on next-generation sequencing (NGS) in ...China. This study was performed to establish a reference system for performance evaluation of BRCA genetic testing and variant interpretation, which includes interpretation rules, reference materials (RMs), and a reference database (RD). BRCA1/2 mutations identified in cell lines and clinical cases were selected to establish RMs. All mutations were detected by NGS and validated by Sanger sequencing. Variant call format files and standard variant data sets were collected and annotated to build the RD. Participant laboratories were invited to validate this reference system. Interpretation rules for BRCA variants in the Chinese population were generated as a standard for BRCA variant interpretation. Mutational analysis demonstrated that BRCA2 mutations (55%) were more common than BRCA1 mutations (45%) in Chinese patients. Eliminating duplicates from 19,886 variants, the RD contained 750 unique BRCA mutations. Most BRCA1/2 mutations in the reference system were pathogenic or likely pathogenic (RMs, 77.5%; RD, 57%). In total, 91 novel pathogenic/likely pathogenic variants were identified in the RD. The reference system can contribute to NGS performance and high-quality interpretation to facilitate clinical decision making. It could also accelerate the development and application of BRCA mutation detection technologies in China.
Multigene panel testing of breast cancer predisposition genes have been extensively conducted in Europe and America, which is relatively rare in Asia however. In this study, we assessed the frequency ...of germline mutations in 40 cancer predisposition genes, including
BRCA1
and
BRCA2
, among a large cohort of Chinese patients with high hereditary risk of BC. From 2015 to 2016, consecutive BC patients from 26 centers of China with high hereditary risk were recruited (n = 937). Clinical information was collected and next‐generation sequencing (NGS) was performed using blood samples of participants to identify germline mutations. In total, we acquired 223 patients with putative germline mutations, including 159 in
BRCA1/2
, 61 in 15 other BC susceptibility genes and 3 in both
BRCA1/2
and non‐
BRCA1/2
gene. Major mutant non‐
BRCA1/2
genes were
TP53
(n = 18),
PALB2
(n = 11),
CHEK2
(n = 6),
ATM
(n = 6) and
BARD1
(n = 5). No factors predicted pathologic mutations in non‐
BRCA1/2
genes when treated as a whole.
TP53
mutations were associated with HER‐2 positive BC and younger age at diagnosis; and
CHEK2
and
PALB2
mutations were enriched in patients with luminal BC. Among high hereditary risk Chinese BC patients, 23.8% contained germline mutations, including 6.8% in non‐
BRCA1/2
genes.
TP53
and
PALB2
had a relatively high mutation rate (1.9 and 1.2%). Although no factors predicted for detrimental mutations in non‐
BRCA1/2
genes, some clinical features were associated with mutations of several particular genes.
What's new?
The prevalence of mutations in breast cancer predisposition genesare not well investigated in Asia. We assessed germline mutations of 40 cancer susceptibility genes in 937 consecutive selected breast cancer patients from 26 centers of China, and discovered 23.8% of participates carried the pathogenic mutation, including 6.8% with mutations in non‐
BRCA1/2
genes, while
TP53
and
PALB2
had a relatively high mutation rates (1.9% and 1.2%).There was no factors predicted for detrimental mutations in non‐
BRCA1/2
genes when treated as a whole.
The bacterial strain Paracoccus denitrificans W12, which could utilize pyridine as its sole source of carbon and nitrogen, was added into a membrane bioreactor (MBR) to enhance the treatment of a ...pharmaceutical wastewater. The treatment efficiencies investigated showed that the removal of chemical oxygen demand, total nitrogen, and total phosphorus were similar between bioaugmented and non-bioaugmented MBRs, however, significant removal of pyridine was obtained in the bioaugmented reactor. When the hydraulic retention time was 60 hr and the influent concentration of pyridine was 250–500 mg/L, the mean effluent concentration of pyridine without adding W12 was 57.2 mg/L, while the pyridine was degraded to an average of 10.2 mg/L with addition of W12. The bacterial community structure of activated sludge during the bioaugmented treatment was analyzed using polymerase chain reaction-denaturing gradient gel electrophoresis (PCR-DGGE). The results showed that the W12 inoculum reversed the decline of microbial community diversity, however, the similarity between bacterial community structure of the original sludge and that of the sludge after bioaugmentation decreased steadily during the wastewater treatment. Sequencing of the DNA recovered from DGGE gel indicated that Flavobacteriaceae sp., Sphingobium sp., Comamonas sp., and Hyphomicrobium sp. were the dominant organisms in time sequence in the bacterial community in the bioaugmented MBR. This implied that the bioaugmentation was affected by the adjustment of whole bacterial community structure in the inhospitable environment, rather than being due solely to the degradation performance of the bacterium added.
Due to lack of systematic reviews, BRCA, DNA Repair Associated (
) mutations in the Chinese population are not completely understood. The following study investigates the prevalence and type of
...mutations in Chinese patients with high hereditary risk of breast cancer (BC). Patients Drwere recruited from 14 cities between October 2015 and February 2016, and were selected based on family and personal medical history.
mutations were analyzed by collecting blood samples from all participants. 437 BC patients were included. A total of seventy-six (17.4%) mutation carriers were identified with no geographic difference. The mutation rate in the early-onset BC patients was lower compared to family history of breast/ovarian cancer (OC), bilateral BC, male BC, BC&OC or meeting ≥2 criteria (9.2 vs. 21.7, 24.0, 22.2, 16.7 and 24.3%, respectively, P=0.007). A total of 61 mutation sites were identified (
32,
29) including 47.5% novel sites and extra 10 variants of uncertain significance. A total of five sites were repeated in more than one unrelated patient. A total of 11 sites were associated with hereditary breast and ovarian cancer syndrome, two of which were confirmed by family pedigrees. Compared with
patients, patients with
mutation tended to be triple-negative BC (P<0.001), whereas patients with
mutation were more likely to be hormone receptor positive BC (P=0.02). The present study provides a general
mutation profile in the Chinese population. The prevalence of
mutation in BC patients with high hereditary risk is lower compared with Western populations. Chinese mutation type is different with Western people, without obvious founder mutation.
PDF
