Novel genotyping algorithms for rare variants significantly improve the accuracy of Applied Biosystems™ Axiom™ array genotyping calls: Retrospective evaluation of UK Biobank array data.

The UK Biobank genotyped about 500k participants using Applied Biosystems Axiom microarrays. Participants were subsequently sequenced by the UK Biobank Exome Sequencing Consortium. Axiom genotyping was highly accurate in comparison to sequencing results, for almost 100,000 variants both directly genotyped on the UK Biobank Axiom array and via whole exome sequencing. However, in a study using the exome sequencing results of the first 50k individuals as reference (truth), it was observed that the positive predictive value (PPV) decreased along with the number of heterozygous array calls per variant. We developed a novel addition to the genotyping algorithm, Rare Heterozygous Adjusted (RHA), to significantly improve PPV in variants with minor allele frequency below 0.01%. The improvement in PPV was roughly equal when comparing to the exome sequencing of 50k individuals, or to the more recent ~200k individuals. Sensitivity was higher in the 200k data. The improved calling algorithm, along with enhanced quality control of array probesets, significantly improved the positive predictive value and the sensitivity of array data, making it suitable for the detection of ultra-rare variants.

Prophetic Granger Causality to infer gene regulatory networks.

We introduce a novel method called Prophetic Granger Causality (PGC) for inferring gene regulatory networks (GRNs) from protein-level time series data. The method uses an L1-penalized regression adaptation of Granger Causality to model protein levels as a function of time, stimuli, and other perturbations. When combined with a data-independent network prior, the framework outperformed all other methods submitted to the HPN-DREAM 8 breast cancer network inference challenge. Our investigations reveal that PGC provides complementary information to other approaches, raising the performance of ensemble learners, while on its own achieves moderate performance. Thus, PGC serves as a valuable new tool in the bioinformatics toolkit for analyzing temporal datasets. We investigate the general and cell-specific interactions predicted by our method and find several novel interactions, demonstrating the utility of the approach in charting new tumor wiring.

A Community Challenge for Inferring Genetic Predictors of Gene Essentialities through Analysis of a Functional Screen of Cancer Cell Lines.

We report the results of a DREAM challenge designed to predict relative genetic essentialities based on a novel dataset testing 98,000 shRNAs against 149 molecularly characterized cancer cell lines. We analyzed the results of over 3,000 submissions over a period of 4 months. We found that algorithms combining essentiality data across multiple genes demonstrated increased accuracy; gene expression was the most informative molecular data type; the identity of the gene being predicted was far more important than the modeling strategy; well-predicted genes and selected molecular features showed enrichment in functional categories; and frequently selected expression features correlated with survival in primary tumors. This study establishes benchmarks for gene essentiality prediction, presents a community resource for future comparison with this benchmark, and provides insights into factors influencing the ability to predict gene essentiality from functional genetic screens. This study also demonstrates the value of releasing pre-publication data publicly to engage the community in an open research collaboration.

Inferring causal molecular networks: empirical assessment through a community-based effort.

It remains unclear whether causal, rather than merely correlational, relationships in molecular networks can be inferred in complex biological settings. Here we describe the HPN-DREAM network inference challenge, which focused on learning causal influences in signaling networks. We used phosphoprotein data from cancer cell lines as well as in silico data from a nonlinear dynamical model. Using the phosphoprotein data, we scored more than 2,000 networks submitted by challenge participants. The networks spanned 32 biological contexts and were scored in terms of causal validity with respect to unseen interventional data. A number of approaches were effective, and incorporating known biology was generally advantageous. Additional sub-challenges considered time-course prediction and visualization. Our results suggest that learning causal relationships may be feasible in complex settings such as disease states. Furthermore, our scoring approach provides a practical way to empirically assess inferred molecular networks in a causal sense.

Genome and proteome annotation using automatically recognized concepts and functional networks.

Many tools have been developed for prediction of the function or disease association of genes and proteins, and this continues to be a highly active area of bioinformatics research. Typically, these methods predict which concepts should be annotated to genes or proteins, using terms from ontologies such as Gene Ontology (GO), largely overlooking other ontologies that are available. Here, we set out to broadly evaluate novel, automatically retrieved, gene-term annotations and identify those concepts of publicly available ontologies that can be predicted using a generalized tool for prediction of annotations. We identified terms that perform better than expected by chance using randomly generated gene sets and show that both manually curated terms in GO and automatically recognized terms can be used to develop reasonable predictive models. In all, we characterize terms in over 250 ontologies and identify more than 127,000 statistically significant terms that can be predicted on human genes.