Reconstructing Causal Biological Networks through Active Learning

Reverse-engineering of biological networks is a central problem in systems biology. The use of intervention data, such as gene knockouts or knockdowns, is typically used for teasing apart causal relationships among genes. Under time or resource constraints, one needs to carefully choose which intervention experiments to carry out. Previous approaches for selecting most informative interventions have largely been focused on discrete Bayesian networks. However, continuous Bayesian networks are of great practical interest, especially in the study of complex biological systems and their quantitative properties. In this work, we present an efficient, information-theoretic active learning algorithm for Gaussian Bayesian networks (GBNs), which serve as important models for gene regulatory networks. In addition to providing linear-algebraic insights unique to GBNs, leading to significant runtime improvements, we demonstrate the effectiveness of our method on data simulated with GBNs and the DREAM4 network inference challenge data sets. Our method generally leads to faster recovery of underlying network structure and faster convergence to final distribution of confidence scores over candidate graph structures using the full data, in comparison to random selection of intervention experiments.     

Regulation of Sensory Neuron Precursor Proliferation by Cyclic GMP-Dependent Protein Kinase

Abstract: Cyclic GMP (cGMP) is a molecular messenger involved in diverse cellular processes. Recently, cGMP-dependent protein kinase (cGK) type II was determined to be a regulator of endochondral ossification and bone growth, identifying a role for cGMP in the regulation of cellular proliferation. Here, we demonstrate the presence of cGK type I (cGKI) in cells of the developing trigeminal ganglia. cGKI occurs in some proliferating precursors as evidenced by double labeling with an antibody to cGKI and 5-bromo-2'-deoxyuridine(BrdU) incorporation. Inhibition of cGKI with KT5823 or Rp -8-(4-chlorophenylthio)-guanosine-3',5'-cyclic monophosphorothioate ( Rp -8-pCPT-cGMPS) in chick embryos results in a 30–40% decrease in trigeminal ganglia cell number, and this effect is independent of nitric oxide synthase (NOS). In addition, inhibition of cGKI with Rp -8-pCPT-cGMPS results in a 60% decrease in BrdU incorporation in the trigeminal ganglia of embryonic day 5 chicks. We find that PC12 cells expressing cGKI proliferate more rapidly and incorporate more BrdU than do control cells. The cGKI inhibitor Rp -8-pCPT-cGMPS decreases proliferation and BrdU incorporation in transfected PC12 cells but has no effect on control cells. The PC12 cells do not express NOS, indicating that this effect is also independent of NOS. Thus, cGKI regulates the proliferation of sensory neurons as a result of activation of a NOS-independent pathway, representing a novel pathway by which the number of sensory neurons is regulated.     

Genome-wide association study of <Emphasis Type="Italic">Mycobacterium avium</Emphasis> subspecies <Emphasis Type="Italic">Paratuberculosis</Emphasis> infection in Chinese Holstein

Background

Paratuberculosis is a contagious, chronic and enteric disease in ruminants, which is caused by Mycobacterium avium subspecies paratuberculosis (MAP) infection, resulting in enormous economic losses worldwide. There is currently no effective cure for MAP infection or a vaccine, it is thus important to explore the genetic variants that contribute to host susceptibility to infection by MAP, which may provide a better understanding of the mechanisms of paratuberculosis and benefit animal genetic improvement. Herein we performed a genome-wide association study (GWAS) to identify genomic regions and candidate genes associated with susceptibility to MAP infection in dairy cattle.

Results

Using Illumina Bovine 50?K (54,609 SNPs) and GeneSeek HD (138,893 SNPs) chips, two analytical approaches were performed, GRAMMAR-GC and ROADTRIPS in 937 Chinese Holstein cows, among which individuals genotyped by the 50?K chip were imputed to HD SNPs with Beagle software. Consequently, 15 and 11 significant SNPs (P?<?5?×?10^??5) were identified with GRAMMAR-GC and ROADTDRIPS, respectively. A total of 10 functional genes were in proximity to (i.e., within 1?Mb) these SNPs, including IL4, IL5, IL13, IRF1, MyD88, PACSIN1, DEF6, TDP2, ZAP70 and CSF2. Functional enrichment analysis showed that these genes were involved in immune related pathways, such as interleukin, T cell receptor signaling pathways and inflammatory bowel disease (IBD), implying their potential associations with susceptibility to MAP infection. In addition, by examining the publicly available cattle QTLdb, a previous QTL for MAP was found to be overlapped with one of regions detected currently at 32.5?Mb on BTA23, where the TDP2 gene was anchored.

Conclusions

In conclusion, we identified 26 SNPs located on 15 chromosomes in the Chinese Holstein population using two GWAS strategies with high density SNPs. Integrated analysis of GWAS, biological functions and the reported QTL information helps to detect positional candidate genes and the identification of regions associated with susceptibility to MAP traits in dairy cattle.

     

Modelling species responses to extreme weather provides new insights into constraints on range and likely climate change impacts for Australian mammals

Conservation of species under climate change relies on accurate predictions of species ranges under current and future climate conditions. To date, modelling studies have focused primarily on how changes in long‐term averaged climate conditions are likely to influence species distributions with much less attention paid to the potential effect of extreme events such as droughts and heatwaves which are expected to increase in frequency over coming decades. In this study we explore the benefits of tailoring predictor variables to the specific physiological constraints of species, or groups of species. We show how utilizing spatial predictors of extreme temperature and water availability (heat‐waves and droughts), derived from high‐temporal resolution, long‐term weather records, provides categorically different predictions about the future (2070) distribution of suitable environments for 188 mammal species across different biomes (from arid zones to tropical environments) covering the whole of continental Australia. Models based on long‐term averages‐only and extreme conditions‐only showed similarly high predictive performance tested by hold‐out cross‐validation on current data, and yet some predicted dramatically different future geographic ranges for the same species under 2070 climate scenarios. Our results highlight the importance of accounting for extreme conditions/events by identifying areas in the landscape where species may cope with average conditions, but cannot persist under extreme conditions known or predicted to occur there. Our approach provides an important step toward identifying the location of climate change refuges and danger zones that goes beyond the current standard of extrapolating long‐term climate averages.     

Expression of endogenous proteins in maize hybrids in a multi-location field trial in India

Genetically modified (GM) crops undergo large scale multi-location field trials to characterize agronomics, composition, and the concentration of newly expressed protein(s) [herein referred to as transgenic protein(s)]. The concentration of transgenic proteins in different plant tissues and across the developmental stages of the plant is considered in the safety assessment of GM crops. Reference or housekeeping proteins are expected to maintain a relatively stable expression pattern in healthy plants given their role in cellular functions. Understanding the effects of genotype, growth stage and location on the concentration of endogenous housekeeping proteins may provide insight into the contribution these factors could have on transgenic protein concentrations in GM crops. The concentrations of three endogenous proteins (actin, elongation factor 1-alpha, and glyceraldehyde 3-phosphate dehydrogenase) were measured in several different maize hybrids grown across multiple field locations over 2 years. Leaf samples were collected from healthy plants at three developmental stages across the growing seasons, and protein concentrations were quantified by indirect enzyme-linked immunosorbent assay (ELISA) for each protein. In general, the concentrations of these three endogenous proteins were relatively consistent across hybrid backgrounds, when compared within one growth stage and location (2–26%CV), whereas the concentrations of proteins in the same hybrid and growth stage across different locations were more variable (12–64%CV). In general, the protein concentrations in 2013 and 2014 show similar trends in variability. Some degree of variability in protein concentrations should be expected for both transgenic and endogenous plant-expressed proteins. In the case of GM crops, the potential variation in protein concentrations due to location effects is captured in the current model of multi-location field testing.