VarBin, a novel method for classifying true and false positive variants in NGS data

Accumulation and regulation of anthocyanins in white and purple Tibetan Hulless Barley (Hordeum vulgare L. var. nudum Hook. f.) revealed by combined de novo transcriptomics and metabolomics

BACKGROUND

Colored barley, which may have associated human health benefits, is more desirable than the standard white variety, but the metabolites and molecular mechanisms underlying seedcoat coloration remain unclear.

RESULTS

Here, the development of Tibetan hulless barley was monitored, and 18 biological samples at 3 seedcoat color developmental stages were analyzed by transcriptomic and metabolic assays in Nierumuzha (purple) and Kunlun10 (white). A total of 41 anthocyanin compounds and 4186 DEGs were identified. Then we constructed the proanthocyanin-anthocyanin biosynthesis pathway of Tibetan hulless barley, including 19 genes encoding structural enzymes in 12 classes (PAL, C4H, 4CL, CHS, CHI, F3H, F3'H, DFR, ANS, ANR, GT, and ACT). 11 DEGs other than ANR were significantly upregulated in Nierumuzha as compared to Kunlun10, leading to high levels of 15 anthocyanin compounds in this variety (more than 25 times greater than the contents in Kunlun10). ANR was significantly upregulated in Kunlun10 as compared to Nierumuzha, resulting in higher contents of three anthocyanins compounds (more than 5 times greater than the contents in Nierumuzha). In addition, 22 TFs, including MYBs, bHLHs, NACs, bZips, and WD40s, were significantly positively or negatively correlated with the expression patterns of the structural genes. Moreover, comparisons of homologous gene sequences between the two varieties identified 61 putative SNPs in 13 of 19 structural genes. A nonsense mutation was identified in the coding sequence of the ANS gene in Kunlun10. This mutation might encode a nonfunctional protein, further reducing anthocyanin accumulation in Kunlun10. Then we identified 3 modules were highly specific to the Nierumuzha (purple) using WGCNA. Moreover, 12 DEGs appeared both in the putative proanthocyanin-anthocyanin biosynthesis pathway and the protein co-expression network were obtained and verified.

CONCLUSION

Our study constructed the proanthocyanin-anthocyanin biosynthesis pathway of Tibetan hulless barley. A series of compounds, structural genes and TFs responsible for the differences between purple and white hulless barley were obtained in this pathway. Our study improves the understanding of the molecular mechanisms of anthocyanin accumulation and biosynthesis in barley seeds. It provides new targets for the genetic improvement of anthocyanin content and a framework for improving the nutritional quality of barley.

The application of deep learning for the classification of correct and incorrect SNP genotypes from whole-genome DNA sequencing pipelines

A downside of next-generation sequencing technology is the high technical error rate. We built a tool, which uses array-based genotype information to classify next-generation sequencing–based SNPs into the correct and the incorrect calls. The deep learning algorithms were implemented via Keras. Several algorithms were tested: (i) the basic, naïve algorithm, (ii) the naïve algorithm modified by pre-imposing different weights on incorrect and correct SNP class in calculating the loss metric and (iii)–(v) the naïve algorithm modified by random re-sampling (with replacement) of the incorrect SNPs to match 30%/60%/100% of the number of correct SNPs. The training data set was composed of data from three bulls and consisted of 2,227,995 correct (97.94%) and 46,920 incorrect SNPs, while the validation data set consisted of data from one bull with 749,506 correct (98.05%) and 14,908 incorrect SNPs. The results showed that for a rare event classification problem, like incorrect SNP detection in NGS data, the most parsimonious naïve model and a model with the weighting of SNP classes provided the best results for the classification of the validation data set. Both classified 19% of truly incorrect SNPs as incorrect and 99% of truly correct SNPs as correct and resulted in the F1 score of 0.21 — the highest among the compared algorithms. We conclude the basic models were less adapted to the specificity of a training data set and thus resulted in better classification of the independent, validation data set, than the other tested models.

Genome Wide Identification of Mutational Hotspots in the Apicomplexan Parasite Neospora caninum and the Implications for Virulence

Neospora caninum is an apicomplexan parasite responsible for neosporosis, a disease causing hind limb paralysis in dogs and abortion in cattle, resulting in substantial economic losses to beef and dairy industries. Marked differences in pathogenicity exist between N. caninum strains suggesting that intrinsic genetic differences exist between them. These differences likely exist in genes expressed during the tachyzoite lifecycle stage which is responsible for the pathogenesis of neosporosis. An improved understanding of these genetic differences is essential to understanding N. caninum virulence, though such knowledge is scarce. Using a variant detection workflow we compared the tachyzoite transcriptomes of two N. caninum strains with different virulence properties: NC-Liverpool (virulent) and NC-Nowra (avirulent). This workflow identified 3130 SNPs and 6123 indels between the strains, and nine markers capturing 30 variants were Sanger sequenced for both strains. Sequencing of these loci was extended to an additional eight strains and subsequent phylogenetic analysis supported a genetic population structure comprised of two major clades with no geographical segregation. Sequence polymorphisms within coding regions of tachyzoite-associated genes were concentrated on chromosomes XI and XII, with 19 distinct tachyzoite-associated SNP hotspot regions identified within coding regions of the N. caninum nuclear genome. The variants were predominantly located in loci associated with protein binding, protein-protein interactions, transcription, and translation. Furthermore, 468 nonsynonymous SNPs identified within protein-coding genes were associated with protein kinase activity, protein binding, protein phosphorylation, and proteolysis. This work may implicate these processes and the specific proteins involved as novel effectors of N. caninum tachyzoite virulence.

Incomplete penetrance for isolated congenital asplenia in humans with mutations in translated and untranslated RPSA exons

Isolated congenital asplenia (ICA) is the only known human developmental defect exclusively affecting a lymphoid organ. In 2013, we showed that private deleterious mutations in the protein-coding region of RPSA, encoding ribosomal protein SA, caused ICA by haploinsufficiency with complete penetrance. We reported seven heterozygous protein-coding mutations in 8 of the 23 kindreds studied, including 6 of the 8 multiplex kindreds. We have since enrolled 33 new kindreds, 5 of which are multiplex. We describe here 11 new heterozygous ICA-causing RPSA protein-coding mutations, and the first two mutations in the 5'-UTR of this gene, which disrupt mRNA splicing. Overall, 40 of the 73 ICA patients (55%) and 23 of the 56 kindreds (41%) carry mutations located in translated or untranslated exons of RPSA. Eleven of the 43 kindreds affected by sporadic disease (26%) carry RPSA mutations, whereas 12 of the 13 multiplex kindreds (92%) carry RPSA mutations. We also report that 6 of 18 (33%) protein-coding mutations and the two (100%) 5'-UTR mutations display incomplete penetrance. Three mutations were identified in two independent kindreds, due to a hotspot or a founder effect. Finally, RPSA ICA-causing mutations were demonstrated to be de novo in 7 of the 23 probands. Mutations in RPSA exons can affect the translated or untranslated regions and can underlie ICA with complete or incomplete penetrance.

Novel PLP1 Mutations Identified With Next-Generation Sequencing Expand the Spectrum of PLP1-Associated Leukodystrophy Clinical Phenotypes

Next-generation sequencing was performed for 2 families with an undiagnosed neurologic disease. Analysis revealed X-linked mutations in the proteolipid protein 1 (PLP1) gene, which is associated with X-linked Pelizaeus-Merzbacher disease and Spastic Paraplegia type 2. In family A, the novel PLP1 missense mutation c.617T>A (p.M206K) was hemizygous in the 2 affected male children and heterozygous in the mother. In family B, the novel de novoPLP1 frameshift mutation c.359_369del (p.G120fs) was hemizygous in the affected male child. Although PLP1 mutations have been reported to cause an increasingly wide range of phenotypes inclusive of the dystonia, spastic paraparesis, motor neuronopathy, and leukodystrophy observed in our patients, atypical features included the cerebrospinal fluid deficiency of neurotransmitter and pterin metabolites and the delayed appearance of myelin abnormalities on neuroimaging studies. Next-generation sequencing studies provided a diagnosis for these families with complex leukodystrophy disease phenotypes, which expanded the spectrum of PLP1-associated leukodystrophy clinical phenotypes.

Utilization of Whole-Exome Next-Generation Sequencing Variant Read Frequency for Detection of Lesion-Specific, Somatic Loss of Heterozygosity in a Neurofibromatosis Type 1 Cohort with Tibial Pseudarthrosis

A subset of neurofibromatosis type 1 patients develop tibial dysplasia, which can lead to pseudarthrosis. The tissue from the tibial pseudarthrosis region commonly has a somatic second hit in NF1: single-nucleotide variants, small deletions, or loss of heterozygosity (LOH). We used exome next-generation sequencing (NGS) variant frequency data (allelic imbalance analysis) to detect somatic LOH in pseudarthrosis tissue from three individuals with clinically and diagnostically confirmed neurofibromatosis type 1, and verified the results with microarray. The variant files were parsed and plotted using python scripts, and the NGS variant frequencies between the affected tissue and blood sample were compared. Individuals without somatic single-nucleotide variants or small insertions/deletions were tested for somatic LOH using the NGS variant allele frequencies. One individual's NGS data indicated no LOH in chromosome 17. The other two individuals demonstrated somatic LOH inclusive of NF1: one had an LOH region of approximately one million bases and Contra (NGS copy number program) indicated a somatic deletion and the other individual had LOH for most of chromosome 17q and Contra indicated no copy number change (microarray data verified this sample as copy neutral somatic LOH). Both LOH and copy number variation detected by NGS data correlated with microarray data, demonstrating the somatic LOH second hit can be detected directly from the NGS data.

Evaluation of somatic mutations in tibial pseudarthrosis samples in neurofibromatosis type 1

BACKGROUND

Tibial pseudarthrosis is associated with neurofibromatosis type 1 (NF1) and there is wide clinical variability of the tibial dysplasia in NF1, suggesting the possibility of genetic modifiers. Double inactivation of NF1 is postulated to be necessary for the development of tibial pseudarthrosis, but tissue or cell of origin of the 'second hit' mutation remains unclear.

METHODS

Exome sequencing of different sections of surgically resected NF1 tibial pseudarthrosis tissue was performed and compared to germline (peripheral blood).

RESULTS

A germline NF1 splice site mutation (c.61-2A>T, p.L21 M68del) was identified from DNA extracted from peripheral blood. Exome sequencing of DNA extracted from tissue removed during surgery of the tibial pseudarthrosis showed a somatic mutation of NF1 (c.3574G>T, p.E1192*) in the normal germline allele. Further analysis of different regions of the tibial pseudarthrosis sample showed enrichment of the somatic mutation in the soft tissue within the pseudarthrosis site and absence of the somatic mutation in cortical bone. In addition, a germline variant in PTPN11 (c.1658C>T, p.T553M), a gene involved in the RAS signal transduction pathway was identified, although the clinical significance is unknown.

CONCLUSIONS

Given that the NF1 somatic mutation was primarily detected in the proliferative soft tissue at the pseudarthrosis site, it is likely that the second hit occurred in mesenchymal progenitors from the periosteum. These results are consistent with a defect of differentiation, which may explain why the mutation is found in proliferative cells and not within cortical bone tissue, as the latter by definition contains mostly mature differentiated osteoblasts and osteocytes.