Classification-oriented dawid skene model for transferring intelligence from crowds to machines

When a crowdsourcing approach is used to assist the classification of a set of items,the main objective is to classify this set of items by aggregating the worker-provided labels.A secondary objective is to assess the workers’skill levels in this process.A classical model that achieves both objectives is the famous Dawid-Skene model.In this paper,we consider a third objective in this context,namely,to learn a classifier that is capable of labelling future items without further assistance of crowd workers.By extending the DawidSkene model to include the item features into consideration,we develop a Classification-Oriented Dawid Skene(CODS)model,which achieves the three objectives simultaneously.The effectiveness of CODS on this three dimensions of the problem space is demonstrated experimentally.

Shared Gene Regulation during Human Somatic Cell Reprogramming

Human induced pluripotent stem （iPS） cells have the ability to differentiate into all somatic cells and to maintain unlimited self- renewal. Therefore, they have great potential in both basic research and clinical therapy for many diseases. To identify potentially universal mechanisms of human somatic cell reprogramming, we studied gene expression changes in three types of cells undergoing reprogramming. The set of 570 genes commonly regulated during induction of iPS cells includes known embryonic stem （ES） cell markers and pluripotency related genes. We also identified novel genes and biological categories which may be related to somatic cell reprogramming. For example, some of the down-regulated genes are predicted targets of the pluripotency microRNA cluster miR302/367, and the proteins from these putative target genes interact with the stem cell pluripotency factor POU5F1 according to our network analysis. Our results identified candidate gene sets to guide research on the mechanisms operating during somatic cell reprogramming.

Variability in pattern and hydrogen isotope composition(δ^(2)H)of long-chain n-alkanes of surface soils and its relations to climate and vegetation characteristics:A meta-analysis

The average chain length(ACL),carbon preference index(CPI),and hydrogen isotope composition(δ^(2)H)of long-chain n-alkanes in sediments have been used to retrieve information about the paleoclimate.Despite their importance as in-between media from leaves to sediments,n-alkanes of surface soils have not been systematically analyzed at large scale.Such an investigation of the spatial variation of n-alkane properties in soil and their dependence on climatic and botanic(e.g.,vegetation type)factors could provide a rationale for a better estimation of the past environment.We synthesized the patterns andδ^(2)H of long-chain n-alkanes in soil(δ^(2)H_(n-alkanes))with regard to vegetation types(cropland,grassland,shrubland,and woodland)and environmental factors using data from peer-reviewed papers.Our results showed that the ACL and CPI of soil C_(27)–C_(33) n-alkanes were not suitable indicators for differentiating vegetation types at large scale;instead,ACL significantly correlated with water conditions such as mean annual precipitation(MAP)and Palmer drought severity index(PDSI),and CPI significantly correlated with temperature without significant influence of vegetation type.The variation(i.e.,standard deviation)of fractionation between theδ^(2)H values in annual precipitation and in soil n-alkanes(ε_(rain-soil))was smaller than that reported in leaves;therefore,soils were better suited to quantifying the general growing conditions of plants at a certain site.The fractionationε_(rain-soil)correlated with climatic conditions as described by the PDSI and relative humidity(RH).This correlation agreed with the change in leaf water enrichment with changing RH taken from the literature and was independent of the vegetation type at large scale.This meta-analysis may provide useful information for the variations of the patterns andδ^(2)H_(n-alkanes) values in surface soils.

RePhine: An Integrative Method for Identification of Drug Response-related Transcriptional Regulators

Transcriptional regulators(TRs)participate in essential processes in cancer pathogenesis and are critical therapeutic targets.Identification of drug response-related TRs from cell line-based compound screening data is often challenging due to low m RNA abundance of TRs,protein modifications,and other confounders(CFs).In this study,we developed a regression-based pharmacogenomic and Ch IP-seq data integration method(Re Phine)to infer the impact of TRs on drug response through integrative analyses of pharmacogenomic and Ch IP-seq data.Re Phine was evaluated in simulation and pharmacogenomic data and was applied to pan-cancer datasets with the goal of biological discovery.In simulation data with added noises or CFs and in pharmacogenomic data,Re Phine demonstrated an improved performance in comparison with three commonly used methods(including Pearson correlation analysis,logistic regression model,and gene set enrichment analysis).Utilizing Re Phine and Cancer Cell Line Encyclopedia data,we observed that Re Phinederived TR signatures could effectively cluster drugs with different mechanisms of action.Re Phine predicted that loss-offunction of EZH2/PRC2 reduces cancer cell sensitivity toward the BRAF inhibitor PLX4720.Experimental validation confirmed that pharmacological EZH2 inhibition increases the resistance of cancer cells to PLX4720 treatment.Our results support that Re Phine is a useful tool for inferring drug response-related TRs and for potential therapeutic applications.The source code for Re Phine is freely available at https://github.com/coexps/Re Phine.