Nanopore sequencing of forensic short tandem repeats using QNome of Qitan Technology.

Devices of nanopore sequencing can be highly portable and of low cost. Thus, nanopore sequencing is promising in in-field forensic applications. Previous investigations have demonstrated that nanopore sequencing is feasible for genotyping forensic short tandem repeats (STRs) by using sequencers of Oxford Nanopore Technologies. Recently, Qitan Technology launched a new portable nanopore sequencer and became the second supplier in the world. Here, for the first time, we assess the QNome (QNome-3841) for its accuracy in nanopore sequencing of STRs and compare with MinION (MinION Mk1B). We profile 54 STRs of 21 unrelated individuals and 2800M standard DNA. The overall accuracy for diploid STRs and haploid STRs were 53.5% (378 of 706) and 82.7% (134 of 162), respectively, by using QNome. The accuracies were remarkably lower than those of MinION (diploid STRs, 84.5%; haploid, 90.7%), with a similar amount of sequencing data and identical bioinformatics analysis. Although it was not reliable for diploid STRs typing by using QNome, the haploid STRs were consistently correctly typed. The majority of errors (58.8%) in QNome-based STR typing were one-repeat deviations of repeat units in the error from true allele, related with homopolymers in repeats of STRs.

CodonBERT: a BERT-based architecture tailored for codon optimization using the cross-attention mechanism.

MOTIVATION: Due to the varying delivery methods of mRNA vaccines, codon optimization plays a critical role in vaccine design to improve the stability and expression of proteins in specific tissues. Considering the many-to-one relationship between synonymous codons and amino acids, the number of mRNA sequences encoding the same amino acid sequence could be enormous. Finding stable and highly expressed mRNA sequences from the vast sequence space using in silico methods can generally be viewed as a path-search problem or a machine translation problem. However, current deep learning-based methods inspired by machine translation may have some limitations, such as recurrent neural networks, which have a weak ability to capture the long-term dependencies of codon preferences. RESULTS: We develop a BERT-based architecture that uses the cross-attention mechanism for codon optimization. In CodonBERT, the codon sequence is randomly masked with each codon serving as a key and a value. In the meantime, the amino acid sequence is used as the query. CodonBERT was trained on high-expression transcripts from Human Protein Atlas mixed with different proportions of high codon adaptation index codon sequences. The result showed that CodonBERT can effectively capture the long-term dependencies between codons and amino acids, suggesting that it can be used as a customized training framework for specific optimization targets. AVAILABILITY AND IMPLEMENTATION: CodonBERT is freely available on https://github.com/FPPGroup/CodonBERT.

A benchmark for evaluation of structure-based online tools for antibody-antigen binding affinity.

The prediction of binding affinity changes caused by missense mutations can elucidate antigen-antibody interactions. A few accessible structure-based online computational tools have been proposed. However, selecting suitable software for particular research is challenging, especially research on the SARS-CoV-2 spike protein with antibodies. Therefore, benchmarking of the mutation-diverse SARS-CoV-2 datasets is critical. Here, we collected the datasets including 1216 variants about the changes in binding affinity of antigens from 22 complexes for SARS-CoV-2 S proteins and 22 monoclonal antibodies as well as applied them to evaluate the performance of seven binding affinity prediction tools. The tested tools' Pearson correlations between predicted and measured changes in binding affinity were between -0.158 and 0.657, while accuracy in classification tasks on predicting increasing or decreasing affinity ranged from 0.444 to 0.834. These tools performed relatively better on predicting single mutations, especially at epitope sites, whereas poor performance on extremely decreasing affinity. The tested tools were relatively insensitive to the experimental techniques used to obtain structures of complexes. In summary, we constructed a list of datasets and evaluated a range of structure-based online prediction tools that will explicate relevant processes of antigen-antibody interactions and enhance the computational design of therapeutic monoclonal antibodies.

THPLM: a sequence-based deep learning framework for protein stability changes prediction upon point variations using pretrained protein language model.

MOTIVATION: Quantitative determination of protein thermodynamic stability is a critical step in protein and drug design. Reliable prediction of protein stability changes caused by point variations contributes to developing-related fields. Over the past decades, dozens of structure-based and sequence-based methods have been proposed, showing good prediction performance. Despite the impressive progress, it is necessary to explore wild-type and variant protein representations to address the problem of how to represent the protein stability change in view of global sequence. With the development of structure prediction using learning-based methods, protein language models (PLMs) have shown accurate and high-quality predictions of protein structure. Because PLM captures the atomic-level structural information, it can help to understand how single-point variations cause functional changes. RESULTS: Here, we proposed THPLM, a sequence-based deep learning model for stability change prediction using Meta's ESM-2. With ESM-2 and a simple convolutional neural network, THPLM achieved comparable or even better performance than most methods, including sequence-based and structure-based methods. Furthermore, the experimental results indicate that the PLM's ability to generate representations of sequence can effectively improve the ability of protein function prediction. AVAILABILITY AND IMPLEMENTATION: The source code of THPLM and the testing data can be accessible through the following links: https://github.com/FPPGroup/THPLM.

Refined Qingkailing protects the in vitro neurovascular unit against oxygen-glucose deprivation and re-oxygenation-induced injury.

Since the proposal of the neurovascular unit (NVU) theory, it has become almost mandatory for neuroprotective medicines against ischaemic stroke (IS) to focus on this unit. Refined Qingkailing (RQKL) is a compound composed of hyodeoxycholic acid, geniposide, baicalin and cholic acid, which has shown great potential in the treatment of IS, but its effect on NVU has not been fully studied. The purpose of this study was to investigate the potential biological pathways that underlie the protective effects of RQKL against NVU damage induced by oxygen-glucose deprivation and re-oxygenation (OGD/R). Using in vitro OGD/R models, we looked into whether RQKL protects the NVU. In order to create an in vitro NVU that resembles IS, we created an OGD/R injury model using primary cultures of brain microvascular endothelial cells, neurons, and astrocytes. Based on our results, we present evidence, for the first time, that RQKL treatment of the injury caused by OGD/R significantly (1) kept the blood brain barrier (BBB) functioning and maintained the architecture of the neurons, (2) mitigated the oxidative stress damage, inflammatory cytokine release, and neuronal death, and (3) upregulated the expression of neurotrophic factors generated from glial cells and the brain in the in vitro model. Therefore, RQKL has a variety of preventive effects against NVU damage caused by OGD/R. RQKL may be a suitable medication for treating IS in a clinical setting.

Emerging effects of non-coding RNA in vascular endothelial cells during strokes.

Vascular and neurological damage are the typical outcomes of ischemic strokes. Vascular endothelial cells (VECs), a substantial component of the blood-brain barrier (BBB), are necessary for normal cerebrovascular physiology. During an ischemic stroke (IS), changes in the brain endothelium can lead to a BBB rupture, inflammation, and vasogenic brain edema, and VECs are essential for neurotrophic effects and angiogenesis. Non-coding RNAs (nc-RNAs) are endogenous molecules, and brain ischemia quickly changes the expression patterns of several non-coding RNA types, such as microRNA (miRNA/miR), long non-coding RNA (lncRNA), and circular RNA (circRNA). Furthermore, vascular endothelium-associated nc-RNAs are important mediators in the maintenance of healthy cerebrovascular function. In order to better understand how VECs are regulated epigenetically during an IS, in this review, we attempted to assemble the molecular functions of nc-RNAs that are linked with VECs during an IS.

Model performance and interpretability of semi-supervised generative adversarial networks to predict oncogenic variants with unlabeled data.

BACKGROUND: It remains an important challenge to predict the functional consequences or clinical impacts of genetic variants in human diseases, such as cancer. An increasing number of genetic variants in cancer have been discovered and documented in public databases such as COSMIC, but the vast majority of them have no functional or clinical annotations. Some databases, such as CiVIC are available with manual annotation of functional mutations, but the size of the database is small due to the use of human annotation. Since the unlabeled data (millions of variants) typically outnumber labeled data (thousands of variants), computational tools that take advantage of unlabeled data may improve prediction accuracy. RESULT: To leverage unlabeled data to predict functional importance of genetic variants, we introduced a method using semi-supervised generative adversarial networks (SGAN), incorporating features from both labeled and unlabeled data. Our SGAN model incorporated features from clinical guidelines and predictive scores from other computational tools. We also performed comparative analysis to study factors that influence prediction accuracy, such as using different algorithms, types of features, and training sample size, to provide more insights into variant prioritization. We found that SGAN can achieve competitive performances with small labeled training samples by incorporating unlabeled samples, which is a unique advantage compared to traditional machine learning methods. We also found that manually curated samples can achieve a more stable predictive performance than publicly available datasets. CONCLUSIONS: By incorporating much larger samples of unlabeled data, the SGAN method can improve the ability to detect novel oncogenic variants, compared to other machine-learning algorithms that use only labeled datasets. SGAN can be potentially used to predict the pathogenicity of more complex variants such as structural variants or non-coding variants, with the availability of more training samples and informative features.

Linking Nonalcoholic Fatty Liver Disease and Brain Disease: Focusing on Bile Acid Signaling.

A metabolic illness known as non-alcoholic fatty liver disease (NAFLD), affects more than one-quarter of the world's population. Bile acids (BAs), as detergents involved in lipid digestion, show an abnormal metabolism in patients with NAFLD. However, BAs can affect other organs as well, such as the brain, where it has a neuroprotective effect. According to a series of studies, brain disorders may be extrahepatic manifestations of NAFLD, such as depression, changes to the cerebrovascular system, and worsening cognitive ability. Consequently, we propose that NAFLD affects the development of brain disease, through the bile acid signaling pathway. Through direct or indirect channels, BAs can send messages to the brain. Some BAs may operate directly on the central Farnesoid X receptor (FXR) and the G protein bile acid-activated receptor 1 (GPBAR1) by overcoming the blood-brain barrier (BBB). Furthermore, glucagon-like peptide-1 (GLP-1) and the fibroblast growth factor (FGF) 19 are released from the intestine FXR and GPBAR1 receptors, upon activation, both of which send signals to the brain. Inflammatory, systemic metabolic disorders in the liver and brain are regulated by the bile acid-activated receptors FXR and GPBAR1, which are potential therapeutic targets. From a bile acid viewpoint, we examine the bile acid signaling changes in NAFLD and brain disease. We also recommend the development of dual GPBAR1/FXR ligands to reduce side effects and manage NAFLD and brain disease efficiently.

Elimination of Foreign Sequences in Eukaryotic Viral Reference Genomes Improves the Accuracy of Virome Analysis.

Widespread in public databases, foreign contaminant sequences pose a substantial obstacle in genomic analyses. Such contamination in viral genome databases is also notorious but more complicated and often causes questionable results in various applications, particularly in virome-based virus detection. Here, we conducted comprehensive screening and identification of the foreign sequences hidden in the largest eukaryotic viral genome collections of GenBank and UniProt using a scrutiny pipeline, which enables us to rigorously detect those problematic viral sequences (PVSs) with origins in hosts, vectors, and laboratory components. As a result, a total of 766 nucleotide PVSs and 276 amino acid PVSs with lengths up to 6,605 bp were determined, which were widely distributed in 39 families with many involving highly public health-concerning viruses, such as hepatitis C virus, Crimean-Congo hemorrhagic fever virus, and filovirus. The majority of these PVSs are genomic fragments of hosts including humans and bacteria. However, they cannot simply be regarded as foreign contaminants, since parts of them are results of natural occurrence or artificial engineering of viruses. Nevertheless, they severely disturb such sequence-based analyses as genome annotation, taxonomic assignment, and virome profiling. Therefore, we provide a clean version of the eukaryotic viral reference data set by the removal of these PVSs, which allows more accurate virome analysis with less time consumed than with other comprehensive databases. IMPORTANCE High-throughput sequencing-based viromics highly depends on reference databases, but foreign contamination is widespread in public databases and often leads to confusing and even wrong conclusions in genomic analysis and viromic profiling. To address this issue, we systematically detected and identified the contamination in the largest viral sequence collections of GenBank and UniProt based on a stringent scrutiny pipeline. We found hundreds of PVSs that are related to hosts, vectors, and laboratory components. By the removal of them, the resulting data set greatly improves the accuracy and efficiency of eukaryotic virome profiling. These results refresh our knowledge of the type and origin of PVSs and also have warning implications for viromic analysis. Viromic practitioners should be aware of these problems caused by PVSs and need to realize that a careful review of bioinformatic results is necessary for a reliable conclusion.

CancerVar: An artificial intelligence-empowered platform for clinical interpretation of somatic mutations in cancer.

Several knowledgebases are manually curated to support clinical interpretations of thousands of hotspot somatic mutations in cancer. However, discrepancies or even conflicting interpretations are observed among these databases. Furthermore, many previously undocumented mutations may have clinical or functional impacts on cancer but are not systematically interpreted by existing knowledgebases. To address these challenges, we developed CancerVar to facilitate automated and standardized interpretations for 13 million somatic mutations based on the AMP/ASCO/CAP 2017 guidelines. We further introduced a deep learning framework to predict oncogenicity for these variants using both functional and clinical features. CancerVar achieved satisfactory performance when compared to several independent knowledgebases and, using clinically curated datasets, demonstrated practical utility in classifying somatic variants. In summary, by integrating clinical guidelines with a deep learning framework, CancerVar facilitates clinical interpretation of somatic variants, reduces manual work, improves consistency in variant classification, and promotes implementation of the guidelines.

Actual mutational research of 19 autosomal STRs based on restricted mutation model and big data.

Short tandem repeat (STR) markers have been widely used in forensic paternity testing and individual identification, but the STR mutation might impact on the forensic result interpretation. Importantly, the STR mutation rate was underestimated due to ignoring the "hidden" mutation phenomenon in most similar studies. Considering this, we use Slooten and Ricciardi's restricted mutation model based on big data to obtain more accurate mutation rates for each marker. In this paper, the mutations of 20 autosomal STRs loci (D3S1358, D1S1656, D13S317, Penta E, D16S539, D18S51, D2S1338, CSF1PO, Penta D, TH01, vWA, D21S11, D6S1043, D7S820, D5S818, TPOX, D8S1179, D12S391, D19S433, and FGA; The restricted model does not include the correction factor of D6S1043, this paper calculates remaining 19 STR loci mutation rates) were investigated in 28,313 (Total: 78,739 individuals) confirmed parentage-testing cases in Chinese Han population. As a result, total 1665 mutations were found in all loci, including 1614 one-steps, 34 two-steps, 8 three-steps, and 9 nonintegral mutations. The loci-specific average mutation rates ranged from 0.00007700 (TPOX) to 0.00459050 (FGA) in trio's and 0.00000000 (TPOX) to 0.00344850 (FGA) in duo's. We analyzed the relationship between mutation rates of the apparent and actual, the trio's and duo's, the paternal and maternal, respectively. The results demonstrated that the actual mutation rates are more than the apparent mostly, and the values of µ1"/µ2"(apparent) are also greater than µ1/µ2 (actual) commonly (µ1", µ1; µ2", µ2 are the mutation rates of one-step and two-step). Therefore, the "hidden" mutations are identified. In addition, the mutations rates of trio's and duo's, the paternal and maternal, exhibit significant difference. Next, those mutation data are used to do a comparison with the studies of other Han populations in China, which present the temporal and regional disparities. Due to the large sample size, some rare mutation events, such as monozygotic (MZ) mutation and "fake four-step mutation", are also reported in this study. In conclusion, the estimation values of actual mutations are obtained based on big data, they can not only provide basic data for the Chinese forensic DNA and population genetics databases, but also have important significance for the development of forensic individual identification, paternity testing and genetics research.

Forensic nanopore sequencing of STRs and SNPs using Verogen's ForenSeq DNA Signature Prep Kit and MinION.

The MinION nanopore sequencing device (Oxford Nanopore Technologies, Oxford, UK) is the smallest commercially available sequencer and can be used outside of conventional laboratories. The use of the MinION for forensic applications, however, is hindered by the high error rate of nanopore sequencing. One approach to solving this problem is to identify forensic genetic markers that can consistently be typed correctly based on nanopore sequencing. In this pilot study, we explored the use of nanopore sequencing for single nucleotide polymorphism (SNP) and short tandem repeat (STR) profiling using Verogen's (San Diego, CA, USA) ForenSeq DNA Signature Prep Kit. Thirty single-contributor samples and DNA standard material 2800 M were genotyped using the Illumina (San Diego, CA, USA) MiSeq FGx and MinION (with R9.4.1 flow cells) devices. With an optimized cutoff for allelic imbalance, all 94 identity-informative SNP loci could be genotyped reliably using the MinION device, with an overall accuracy of 99.958% (1 error among 2926 genotypes). STR typing was notably error prone, and its accuracy was locus dependent. We developed a custom-made bioinformatics workflow, and finally selected 13 autosomal STRs, 14 Y-STRs, and 4 X-STRs showing high consistency between nanopore and Illumina sequencing among the tested samples. These SNP and STR loci could be candidates for panel design for forensic analysis based on nanopore sequencing.

Antibacterial Cyclic Tripeptides from Antarctica-Sponge-Derived Fungus Aspergillus insulicola HDN151418.

Three new aspochracin-type cyclic tripeptides, sclerotiotides M-O (1-3), together with three known analogues, sclerotiotide L (4), sclerotiotide F (5), and sclerotiotide B (6), were obtained from the ethyl acetate extract of the fungus Aspergillus insulicola HDN151418, which was isolated from an unidentified Antarctica sponge. Spectroscopic and chemical approaches were used to elucidate their structures. The absolute configuration of the side chain in compound 4 was elucidated for the first time. Compounds 1 and 2 showed broad antimicrobial activity against a panel of pathogenic strains, including Bacillus cereus, Proteus species, Mycobacterium phlei, Bacillus subtilis, Vibrio parahemolyticus, Edwardsiella tarda, MRCNS, and MRSA, with MIC values ranging from 1.56 to 25.0 µM.

A 472-SNP panel for pairwise kinship testing of second-degree relatives.

Kinship testing based on genetic markers, as forensic short tandem repeats (STRs) and single nucleotide polymorphisms (SNPs), has valuable practical applications. Paternity and first-degree relationship can be accurately identified by current commonly-used forensic STRs and reported SNP markers. However, second-degree and more distant relationships remain challenging. Although ∼105-106 SNPs can be used to estimate relatedness of higher degrees, genome-wide genotyping and analysis may be impractical for forensic use. With rapid growth of human genome data sets, it is worthwhile to explore additional markers, especially SNPs, for kinship analysis. Here, we reported an autosomal SNP panel consisted of 342 SNP selected from >84 million SNPs and 131 SNPs from previous systems. We genotyped these SNPs in 136 Chinese individuals by multiplex amplicon Massively Parallel Sequencing, and performed pairwise gender-independent kinship testing. The specificity and sensitivity of these SNPs to distinguish second-degree relatives and the unrelated was 99.9% and 100%, respectively, compared with 53.7% and 99.9% of 19 commonly-used forensic STRs. Moreover, the specificity increased to 100% by the combined use of these STRs and SNPs. The 472-SNP panel could also greatly facilitate the discrimination among different relationships. We estimated that the power of ∼6.45 SNPs were equivalent to one forensic STR in the scenario of 2nd-degree relative pedigree. Altogether, we proposed a panel of 472 SNP markers for kinship analysis, which could be important supplementary of current forensic STRs to solve the problem of second-degree relative testing.