Background: Multiple studies have shown the utility of transcriptome-wide RNA-seq profiles as features for machine learning-based prediction of response to chemotherapy in cancer. While tumor transcriptome profiles are publicly available for thousands of tumors for many cancer types, a relatively modest number of tumor profiles are clinically annotated for response to chemotherapy. The paucity of labeled examples and the high dimension of the feature data limit performance for predicting therapeutic response using fully-supervised classification methods. Recently, multiple studies have established the utility of a deep neural network approach, the variational autoencoder (VAE), for generating meaningful latent features from original data. Here, we report the first study of a semi-supervised approach using VAE-encoded tumor transcriptome features and regularized gradient boosted decision trees (XGBoost) to predict chemotherapy drug response for five cancer types: colon, pancreatic, bladder, breast, and sarcoma.Results: We found: (1) VAE-encoding of the tumor transcriptome preserves the cancer type identity of the tumor, suggesting preservation of biologically relevant information; and (2) as a feature-set for supervised classification to predict response-to-chemotherapy, the unsupervised VAE encoding of the tumor’s gene expression profile leads to better area under the receiver operating characteristic curve and area under the precision-recall curve classification performance than the original gene expression profile or the PCA principal components or the ICA components of the gene expression profile, in four out of five cancer types that we tested.Conclusions: Given high-dimensional “omics” data, the VAE is a powerful tool for obtaining a nonlinear low-dimensional embedding; it yields features that retain biological patterns that distinguish between different types of cancer and that enable more accurate tumor transcriptome-based prediction of response to chemotherapy than would be possible using the original data or their principal components.

The growing availability of biomedical data offers vast potential to improve human health, but the complexity and lack of integration of these datasets often limit their utility. To address this, the Biomedical Data Translator Consortium has developed an open‐source knowledge graph–based system—Translator—designed to integrate, harmonize, and make inferences over diverse biomedical data sources. We announce here Translator's initial public release and provide an overview of its architecture, standards, user interface, and core features. Translator employs a scalable, federated, knowledge graph framework for the integration of clinical, genomic, pharmacological, and other biomedical knowledge sources, enabling query retrieval, inference, and hypothesis generation. Translator's user interface is designed to support the exploration of knowledge relationships and the generation of insights, without requiring deep technical expertise and gradually revealing more detailed evidence, provenance, and confidence information, as needed by a given user. To demonstrate Translator's application and impact, we highlight features of the user interface in the context of three real‐world use cases: suggesting potential therapeutics for patients with rare disease; explaining the mechanism of action of a pipeline drug; and screening and validating drug candidates in a model organism. We discuss strengths and limitations of reasoning within a largely federated system and the need for rich concept modeling and deep provenance tracking. Finally, we outline future directions for enhancing Translator's functionality and expanding its data sources. Translator represents a significant step forward in making complex biomedical knowledge more accessible and actionable, aiming to accelerate translational research and improve patient care.

Feline injection-site sarcoma (FISS), an aggressive iatrogenic subcutaneous malignancy, is challenging to manage clinically and little is known about the molecular basis of its pathogenesis. Tumor transcriptome profiling has proved valuable for gaining insights into the molecular basis of cancers and for identifying new therapeutic targets. Here, we report the first study of the FISS transcriptome and the first cross-species comparison of the FISS transcriptome with those of anatomically similar soft-tissue sarcomas in dogs and humans. Using high-throughput short-read paired-end sequencing, we comparatively profiled FISS tumors vs. normal tissue samples as well as cultured FISS-derived cell lines vs. skin-derived fibroblasts. We analyzed the mRNA-seq data to compare cancer/normal gene expression level, identify biological processes and molecular pathways that are associated with the pathogenesis of FISS, and identify multimegabase genomic regions of potential somatic copy number alteration (SCNA) in FISS. We additionally conducted cross-species analyses to compare the transcriptome of FISS to those of soft-tissue sarcomas in dogs and humans, at the level of cancer/normal gene expression ratios. We found: (1) substantial differential expression biases in feline orthologs of human oncogenes and tumor suppressor genes suggesting conserved functions in FISS; (2) a genomic region with recurrent SCNA in human sarcomas that is syntenic to a feline genomic region of probable SCNA in FISS; and (3) significant overlap of the pattern of transcriptional alterations in FISS with the patterns of transcriptional alterations in soft-tissue sarcomas in humans and in dogs. We demonstrated that a protein, BarH-like homeobox 1 (BARX1), has increased expression in FISS cells at the protein level. We identified 11 drugs and four target proteins as potential new therapies for FISS, and validated that one of them (GSK-1059615) inhibits growth of FISS-derived cells in vitro. (1) Window-based analysis of mRNA-seq data can uncover SCNAs. (2) The transcriptome of FISS-derived cells is highly consistent with that of FISS tumors. (3) FISS is highly similar to soft-tissue sarcomas in dogs and humans, at the level of gene expression. This work underscores the potential utility of comparative oncology in improving understanding and treatment of FISS.

Background:Patients with arterial perinatal stroke often suffer long-term motor sequelae, difficulties in language, social development, and behaviour as well as epilepsy. Despite homogeneous lesions, long-term behavioural and cognitive outcomes are variable and unpredictable. Sleep-related epileptic encephalopathies can occur after early brain injury and are associated with global developmental delays. We hypothesized that sleep-potentiated epileptiform abnormalities are associated with worse developmental outcomes after perinatal stroke.Methods:Participants were identified from a population-based cohort (Alberta Perinatal Stroke Project). Inclusion criteria were magnetic resonance imaging–confirmed arterial perinatal stroke, age 4 to 18 years, electroencephalogram (EEG) including sleep, and comprehensive neuropsychological evaluation. Sleep-related EEG abnormalities were categorized by an epileptologist blinded to the cognitive outcome. Associations between EEG classification and neuropsychological outcomes were explored (ttests, Bonferroni correction for multiple comparisons).Results:Of 128 potentially eligible participants, 34 (53% female) had complete EEG (mean age, 8.1 years; range, 0.2-16.4) and neuropsychology testing (mean age, 9.8 years; range 4.4-16.7). Twelve (35%) were classified as having electrical status epilepticus in sleep. Patients with abnormal EEGs were more likely to have statistically worse scores when corrected for multiple comparisons, in receptive language (median, 1st percentile; IQR 1-7th percentile; p<0.05), and externalizing behaviours (median, 82nd percentile; IQR, 79-97th percentile; p<0.05).Conclusions:Developmental outcome in language and behaviour in children with arterial perinatal stroke is associated with electrical status epilepticus in sleep. Increased screening with sleep EEG is suggested, whereas further studies are necessary to determine if treatment of EEG abnormalities can improve outcome.

Background: Evidence-based medicine has shown that successful vaginal birth after cesarean (VBAC) is associated with fewer complications than an elective repeat cesarean. Although spontaneous vaginal births and reductions in cesarean delivery (CD) rates have been advocated, the risk factors for VBAC complications remain unclear and failed trials of labor (TOL) can lead to adverse pregnancy outcomes.Methods: To construct an antepartum predictive scoring model for VBAC. Retrospective analysis of charts from 1062 women who underwent TOL at no less than 28 gestational weeks with vertex singletons and no more than one prior CD.Results: We constructed our scoring model based on the following variables: maternal age, previous vaginal delivery, interdelivery interval (time between prior cesarean and the following delivery), presence of prior cesarean TOL, dystocia as prior CD indication, intertuberous diameter, maternal predelivery body mass index, gestational age at delivery, estimated fetal weight, and hypertensive disorders. Previous vaginal delivery was the most influential variable. The nomogram showed an area under the curve of 77.7% (95% confidence interval, 73.8–81.5%; sensitivity, 78%; specificity, 70%; cut-off, 13 points). The Kappa value to judge the consistency of the results between the predictive model and the actual results was 0.71(95% confidence interval 0.65–0.77) indicating strong consistency. We used the cut-off to divide the VBAC women into two groups according to the success of the TOL. The maternal and neonatal outcomes such as labor time, number of deliveries by midwives, postpartum hemorrhage, uterine rupture, neonatal asphyxia, puerperal infection were significantly different between the two groups.Conclusions: Our predictive scoring model incorporates easily ascertainable variables and can be used to personalize antepartum counselling for successful TOLs after cesareans.

We previously reported on CERENKOV, an approach for identifying regulatory single nucleotide polymorphisms (rSNPs) that is based on 246 annotation features. CERENKOV uses the xgboost classifier and is designed to be used to find causal noncoding SNPs in loci identified by genome-wide association studies (GWAS). We reported that CERENKOV has state-of-the-art performance (by two traditional measures and a novel GWAS-oriented measure, AVGRANK) in a comparison to nine other tools for identifying functional noncoding SNPs, using a comprehensive reference SNP set (OSU17, 15,331 SNPs). Given that SNPs are grouped within loci in the reference SNP set and given the importance of the data-space manifold geometry for machine-learning model selection, we hypothesized that within-locus inter-SNP distances would have class-based distributional biases that could be exploited to improve rSNP recognition accuracy. We thus defined an intralocus SNP “radius” as the average data-space distance from a SNP to the other intralocus neighbors, and explored radius likelihoods for five distance measures. We expanded the set of reference SNPs to 39,083 (the OSU18 set) and extracted CERENKOV SNP feature data. We computed radius empirical likelihoods and likelihood densities for rSNPs and control SNPs, and found significant likelihood differences between rSNPs and control SNPs. We fit parametric models of likelihood distributions for five different distance measures to obtain ten log-likelihood features that we combined with the 248-dimensional CERENKOV feature matrix. On the OSU18 SNP set, we measured the classification accuracy of CERENKOV with and without the new distance-based features, and found that the addition of distance-based features significantly improves rSNP recognition performance as measured by AUPVR, AUROC, and AVGRANK. Along with feature data for the OSU18 set, the software code for extracting the base feature matrix, estimating ten distance-based likelihood ratio features, and scoring candidate causal SNPs, are released as open-source software CERENKOV2. Accounting for the locus-specific geometry of SNPs in data-space significantly improved the accuracy with which noncoding rSNPs can be computationally identified.