Archive of Story of Palaeobotany Series 196-199 古植物学的故事( 196 ): 古植物学博士:毕业就业愁断肠,改行改志放眼量 Story of Palaeobotany Series (No.196) Chinese palaeobotany students with lots of worry for study and jobs (in Chinese) http://blog.sciencenet.cn/blog-225931-686785.html archive of story of-palaeobotany-ser-196.pdf ----------- 古植物学的故事 ( 第 197 期 ) 《英国专辑》 ( 补充之一 ) 我所认识的第一位英国古植物学家 -- Dianne Edwards ( FRS, 1942-- ) Story of Palaeobotany Series (197) Dianne Edwards (FRS, 1942--), the first British palaeobotanist I have known (The 1 st Addendum to the Special Issue for British Palaeobotany) http://blog.sciencenet.cn/blog-225931-691304.html archive of story of palaeobotany ser-197.pdf -------- 古植物学的故事(第 198 期) 关于中国的古植物学术社团(之九) 中国植物学会古植物学分会怎么办? Story of Palaeobotany Series (No.198) Palaeobotanical community in China Two state-run palaeobotanical associations in China (partIX) Where is the Palaeobotanical Association of Botanical Society of China (BSC) going next? http://blog.sciencenet.cn/blog-225931-697567.html archive of story of palaeobotany ser-198.pdf ----- 古植物学的故事 199 期 以古植物学为例谈建立中国国家自然历史博物馆的必要性与紧迫性(提纲) http://blog.sciencenet.cn/blog-225931-702778.html Story of Palaeobotany Series (No.199) How to promote the rapid rise of Chinese palaeobotany in the round? (Part XVIII) Palaeobotany is employed to illustrate the necessity and urgency of establishing the National Museum of Natural History of China (NMNHC) archive of story of palaeobotany ser-199.pdf ================ Refer to the catalogue: The bilingual Catalogue of the Story of Paeobotany Series (Issues 0 to199) http://blog.sciencenet.cn/blog-225931-708772.html 2013-7-1707:27
Dear Colleague The Urban Studies Foundation is pleased to announce its Seminar Series Competition, which aims to support the generation of internationally excellent research in all areas of urban studies. In the 2013 competition, we seek to support up to three seminar series, in each case up to a maximum of £20,000. Applications to the Competition will be received in two stages and initial applications should be submitted no later than 1 November 2013. We welcome proposals from academics working in any country and all proposals should be clearly international in coverage and scope. Proposals should be in-scope in relation to the interests of the Urban Studies Journal academic community. Further information and the short Initial Application Form are available to download at the Foundation’s website at www.urbanstudiesfoundation.org and enquiries should be made to Ruth Harkin (ruth.harkin@glasgow.ac.uk) in the first instance. Yours sincerely, Professor Joanne Sharp Director Urban Studies Foundation Urban Studies Foundation is a registered charity, reference number SC039937
Introduction to Reading GPS Time Series Plots http://cws.unavco.org:8080/cws/modules/readingGPStimeseries This activity provides a brief introduction to GPS and provides a student activity to practice creating and reading time series plots with simplified GPS data. Students graph how a tectonic plate (and the GPS unit attached to it) has moved over a five year time period by moving a GPS model across a North-East coordinate graph. The following files are needed to complete this activity: Introduction to Reading GPS Time Series Plots (v. May 2011) This presentation provides a brief overview of how GPS can be used to study plate motion, earthquakes, and seismic waves using the Mw=9.0 Tohoku, Japan Earthquake as an example; a brief summary of the development of thought about the shape of the Earth and Geodesy, how GPS works, and step by step how to read GPS time series plots, a brief assessment of learning, and practice with a simple dataset with Iceland as an example. Student Worksheet Reading Time Series Plots (v. May 2011) ; in Spanish v. 2008 Tenths of a Year to Months of a Year Scale Reading Time Series Plots Reference Key ; in Spanish Teacher Guide and Answer Key (v. May 2011) ; After this activity, students explore authentic GPS data; see: Exploring Plate Motion and Deformation in California Using GPS Time Series Plots Using GPS data to visualize plate tectonics in the Pacific Northwest Real GPS Data: Surprising Discoveries in Iceland Additional Resources: The activity: Introduction to Reading Time Series Plots and activity: Exploring Plate Motion and Deformation in California Using GPS Time Series Plots have been tied together in a single Lesson Plan (v.June 2008) This lesson plan also includes an answer key. Draft Versions of this activity
http://a-little-book-of-r-for-time-series.readthedocs.org/en/latest/src/timeseries.html#plotting-time-series http://a-little-book-of-r-for-biomedical-statistics.readthedocs.org/en/latest/#welcome-to-a-little-book-of-r-for-biomedical-statistics Welcome to a Little Book of R for Multivariate Analysis! ¶ By Avril Coghlan , Wellcome Trust Sanger Institute, Cambridge, U.K. Email: alc@sanger.ac.uk This is a simple introduction to multivariate analysis using the R statistics software. There is a pdf version of this booklet available at: https://media.readthedocs.org/pdf/little-book-of-r-for-multivariate-analysis/latest/little-book-of-r-for-multivariate-analysis.pdf . If you like this booklet, you may also like to check out my booklet on using R for biomedical statistics, http://a-little-book-of-r-for-biomedical-statistics.readthedocs.org/ , and my booklet on using R for time series analysis, http://a-little-book-of-r-for-time-series.readthedocs.org/ . Contents: How to install R Introduction to R Installing R How to check if R is installed on a Windows PC Finding out what is the latest version of R Installing R on a Windows PC How to install R on non-Windows computers (eg. Macintosh or Linux computers) Installing R packages How to install an R package How to install a Bioconductor R package Running R A brief introduction to R Links and Further Reading Acknowledgements Contact License Using R for Multivariate Analysis Multivariate Analysis Reading Multivariate Analysis Data into R Plotting Multivariate Data A Matrix Scatterplot A Scatterplot with the Data Points Labelled by their Group A Profile Plot Calculating Summary Statistics for Multivariate Data Means and Variances Per Group Between-groups Variance and Within-groups Variance for a Variable Between-groups Covariance and Within-groups Covariance for Two Variables Calculating Correlations for Multivariate Data Standardising Variables Principal Component Analysis Deciding How Many Principal Components to Retain Loadings for the Principal Components Scatterplots of the Principal Components Linear Discriminant Analysis Loadings for the Discriminant Functions Separation Achieved by the Discriminant Functions A Stacked Histogram of the LDA Values Scatterplots of the Discriminant Functions Allocation Rules and Misclassification Rate Links and Further Reading Acknowledgements Contact License Acknowledgements Thank you to Noel O’Boyle for helping in using Sphinx, http://sphinx.pocoo.org , to create this document, and github, https://github.com/ , to store different versions of the document as I was writing it, and readthedocs, http://readthedocs.org/ , to build and distribute this document. Contact I will be very grateful if you will send me ( Avril Coghlan ) corrections or suggestions for improvements to my email address alc@sanger.ac.uk License The content in this book is licensed under a Creative Commons Attribution 3.0 License .
http://rtutorialseries.blogspot.com/ Good teaching material. Another example for repeated measures ANOVA(This tutorial works for my case) http://ww2.coastal.edu/kingw/statistics/R-tutorials/repeated.html Two Way Mixed Factorial Designs Clean out your workspace, and then enter the data for this analysis bycopying and pasting this script into your R Console... ### begin copying with this line # create the data vectors for the four conditions pre.chicken=c(18,13,18,15,22,32,31,24,15) post.chicken=c(15,13,17,15,24,31,31,25,17) pre.pasta=c(17,30,18,13,23,27,27,24,23) post.pasta=c(19,31,18,13,24,27,26,28,26) # create the response variable vector SSS=c(pre.chicken,post.chicken,pre.pasta,post.pasta) # create the test condition labels test=rep(c(pre,post,pre,post),c(9,9,9,9)) # create the diet condition labels diet=rep(c(chicken,pasta),c(18,18)) # create the subject labels subj.chicken=rep(c(A,B,C,D,E,F,G,H,I),2) subj.pasta=rep(c(J,K,L,M,N,O,P,Q,R),2) # create the subject vector subject=c(subj.chicken,subj.pasta) # create the dataframe hill=data.frame(subject,diet,test,SSS) # clean up rm(pre.chicken,post.chicken,pre.pasta,post.pasta) rm(SSS,test,diet,subj.chicken,subj.pasta,subject) ### end copying with this line You should now have a data frame called hill in your workspace... ls() groceries groceries2 hill str(hill)'data.frame': 36 obs. of 4 variables: $ subject: Factor w/ 18 levels A,B,C,D,..: 1 2 3 4 5 6 7 8 9 1 ... $ diet : Factor w/ 2 levels chicken,pasta: 1 1 1 1 1 1 1 1 1 1 ... $ test : Factor w/ 2 levels post,pre: 2 2 2 2 2 2 2 2 2 1 ... $ SSS : num 18 13 18 15 22 32 31 24 15 15 ... These data were collected by Ben Hill (by now, Ben Hill, Ph.D.) as part of hissenior research project conducted in our department during the Fall 1999semester. His interest was in the role of brain serotonin in creating sensationseeking behavior. Since he couldn't manipulate brain serotonin directly (ourIRB would not permit it!), he chose to do so by way of diet, relying on thefinding that brain serotonin is elevated after a meal rich in carbohydrates.Therefore, he had 18 subjects come to the lab. All subjects were given theSensation Seeking Survey. Half the subjects were then given an evening mealconsisting primarily of chicken, while the other half were given a mealconsisting primarily of pasta. An hour later, the Sensation Seeking Survey wasreadministered. In the data frame, subject identifies the subject so thathis or her pre and post eating scores can be kept paired for the analysis,diet identifies the type of meal the subject was given, test contains thepre-post meal information, and SSS contains the survey scores. In this experiment, we have two explanatory variables, diet, which is abetween-subjects variable, and test, which is a within-subjects variable.Thus, in lingo familiar to social scientists, the design is called a 2 × 2 mixed factorial design with repeated measures on test.It's worth making sure you understand how the data frame is structured, so let'shave a look... hill subject diet test SSS 1 A chicken pre 18 2 B chicken pre 13 3 C chicken pre 18 4 D chicken pre 15 5 E chicken pre 22 6 F chicken pre 32 7 G chicken pre 31 8 H chicken pre 24 9 I chicken pre 15 10 A chicken post 15 11 B chicken post 13 12 C chicken post 17 13 D chicken post 15 14 E chicken post 24 15 F chicken post 31 16 G chicken post 31 17 H chicken post 25 18 I chicken post 17 19 J pasta pre 17 20 K pasta pre 30 21 L pasta pre 18 22 M pasta pre 13 23 N pasta pre 23 24 O pasta pre 27 25 P pasta pre 27 26 Q pasta pre 24 27 R pasta pre 23 28 J pasta post 19 29 K pasta post 31 30 L pasta post 18 31 M pasta post 13 32 N pasta post 24 33 O pasta post 27 34 P pasta post 26 35 Q pasta post 28 36 R pasta post 26 Note that all the SSS scores are in ONE column! The other columns completelyidentify the experimental conditions associated with that score, includingwhich subject it came from. By way of data summary... with(hill, tapply(SSS, list(diet,test), mean)) post pre chicken 20.88889 20.88889 pasta 23.55556 22.44444 Drat! There goes R arranging our factor levels in alphabetical order again,making our means table look backwards. Let's pretty that up... hill$test = factor(hill$test, levels=c(pre,post)) with(hill, tapply(SSS, list(diet,test), mean)) pre post chicken 20.88889 20.88889 pasta 22.44444 23.55556 Better! A nice graph might be in order as well... with(hill, boxplot(SSS ~ diet + test)) # output not shown with(hill, boxplot(SSS ~ test + diet)) # compare this one with the last one title(main=Ben Hill's SSS Data) title(ylab=SSS Scores) The SSS scores in the pasta group did go up as hypothesized, but there is a lotof within groups variability there, too. Will the effect turn out to bestatistically significant? Once again, the trick is in getting the model formula correct. In this case,we have two explanatory variables, and we want to see all possible main effectsand interactions. The test variable is within subjects... aov.out = aov(SSS ~ diet * test + Error(subject/test), data=hill) summary(aov.out) Error: subject Df Sum Sq Mean Sq F value Pr(F) diet 1 40.11 40.11 0.5094 0.4857 Residuals 16 1259.78 78.74 Error: subject:test Df Sum Sq Mean Sq F value Pr(F)test 1 2.7778 2.7778 2.1739 0.1598 diet:test 1 2.7778 2.7778 2.1739 0.1598 Residuals 16 20.4444 1.2778 The effect we're looking for would be shown by the diet × testinteraction. And sadly, it's not significant. Other Designs For reference, here are model formulae for a couple other common designs... Two factor design with repeated measures on both factors: DV ~ IV1 * IV2 + Error(subject/(IV1*IV2)) Three factor mixed design with repeated measures on IV2 and IV3: DV ~ IV1 * IV2 * IV3 + Error(subject/(IV2*IV3)) And so on.
暑假已结束了,感觉假期总体效率不高,中间忙了些别的,还略微看了些心理学 总结下暑假自己看的专业书(不包括读物和论文等),其实很多以前略有接触,没有全部看。 Matlab Getting Started,Matlab User's Guide(这两个是自带的,全当总结用) Networks An Introduction M. E. J. Newman 2010 .pdf(这是复杂网络的入门) Mathematical.Statistics.and.Data.Analysis,.3ed,.Rice,.pdf(增补数理统计部分) Analysis of Financial Time Series 3rd Edition.Tsay 2010 .pdf(时间序列中级教材) Elementary Differential Equations.sixth edition C.Henry Edwards.pdf(增补指数、线性系统、边界值等) Differential equations, dynamical systems and an introduction to chaos.pdf(增补动力系统概念) Dynamical Systems(Jost).pdf(这个太难了,后来发现是数学专业的专题,看了一半不了了之) 现在正在看或马上准备看的 Nonlinear Systems - Analysis, Stability, and Control (Shankar Sastry).pdf Synchronization A universal concept in nonlinear sciences.Arkady Pikovsky, Michael Rosenblum and Jürgen Kurths.pdf Nonlinear Time Series Analysis.second edition,holger kantz and thomas schreiber.pdf 按照之前戴老师的书单,还有好多基础书要看啊! 哈哈,下次再总结下到目前为止数学基础看了多少
终于差不多看完Analysis of Financial Time Series,中间虽干了些别的,但总共大概花了2-3周才看完,暑假效率不高啊! 书的第4、5、9、12章没太看懂。 4 Nonlinear Models and Their Applications 很多东西没讲清楚,准备用下书好好增补,这个以后是重点 Nonlinear Time Series Analysis, Holger Kantz 5 High-Frequency Data Analysis and Market Microstructure直接在分析股票短线操作,没有金融相关基础 9 Principal Component Analysis and Factor Models 在分析个股的行业相关性等,最后的例子尽是分析必和必拓和淡水河谷的股票 12 Markov Chain Monte Carlo Methods with Applications 主要在介绍贝叶斯方法。本科课本只字未提贝叶斯学派,且比频域学派理解起来要困难。虽然之前暑假自学统计学时书上有专门介绍,但还是彻底看不懂,感觉区区60页很多东西没讲清楚。 接下来准备看看R语言,做做课后习题,作者的个人主页上还有不少补充材料,可以继续学习的。
UHawaii Manoa Campus Events Calendar Pacific Asian Lecture Series: Sino-Capitalism - China's Reemergence and more May 29, 4:15pm - 5:30pm Manoa Campus, Shidler College of Business, Faculty Lounge BusAd D-207 http://www.hawaii.edu/calendar/manoa/2012/05/29/18490.html More on Sino-Capitalism: http://www.eastwestcenter.org/news-center/east-west-wire/adapting-rise-sino-capitalism
2013_4EN.pdf Call for papers Special Section on Advances and Applications in Nonlinear Time Series Methods The Institute of Electronics, Information and Communication Engineers (IEICE) We are pleased to announce a Special Section (SS) of the IEICE’s new journal, “Nonlinear Theory and Its Applications, IEICE,” to be published in April 2013. The major part of this SS will focus on recent progress on both methodological advances in and novel applications of nonlinear time series methods. The topics of interest within the scope of this SS include, but are not limited to, the following areas: - Nonlinear time series methods for scalar time series: embedding, invariants and non-stationarity - Embedding and reconstruction - Multivariate recordings - causality and correlation - Novel network and recurrence based methods - Complex networks, complex systems and time series - Applications: climate, physiology and medicine, finance, physical processes Papers submitted to this SS will be peer-reviewed under the handling of the editorial committee of the SS. The deadline of the paper submission is July 10, 2012. Prospective authors are requested to carefully follow the submission process described below: (1) Submit a paper using the IEICE Web site ( https://review.ieice.org/regist_common_e.aspx?society=NOLTA ). Authors should choose " Nonlinear Time Series Methods" as a "Type of Issue (Section)/ Transactions" on the line screen. Please do not choose " ". (2) Send "Copyright Transfer and Article Charge Agreement" and "Confirmation Sheet of Manuscript Registration" by1. E-mail, FAX or postal mail to the following address: Tomomichi Nakamura Graduate School of Simulation Studies, University of Hyogo 7-1-28 Minatojima-minamimachi, Chuo-ku, Kobe, Hyogo 650-0047, Japan E-mail: nolta.timeseries@gmail.com Fax: +81-78-303-1985 (please attach a cover sheet indicating the sender's name) The article charges until 30 pages: One of the authors is a member of IEICE 60,000 JPY Otherwise 65,000 JPY Please do not forget to send "Copyright Transfer and Article Charge Agreement" and "Confirmation Sheet of Manuscript Registration" by July 10, 2012. We cannot start the review process without them, even if we receive their manuscripts. For additional guidelines on manuscript preparation, please visit the following site: http://www.nolta.ieice.org/authors.html Please note that if accepted, authors are requested to pay for the article charges covering partial costs of publications, page charge cannot normally be waived. Both members and nonmembers of IEICE are invited to submit manuscript. However, we recommend that authors unaffiliated with IEICE apply for membership: http://www.ieice.org/eng/member/OM-appli.html Editorial Committee of the Special Section Guest Editors Michael Small (The University of Western Australia) Tomomichi Nakamura (University of Hyogo) Advisory Editor Yoshifumi Nishio (Tokushima University)
Time Series Econometrics http://www.ect.uni-bonn.de/studium/wintersemester-10-11/time-series-econometrics Prof. Dr. Jörg Breitung Sommersemester 2010 Inhalt: Mit Hilfe der multivariaten Zeitreihenanalyse können dynamische Zusammenhänge zwischen makroökonomischen Zeitreihen beschrieben werden. Die Analyse der langfristigen Beziehungen zwischen den Zeitreihen kann mit Hilfe der Kointegrationsanalyse untersucht werden. Solche Kointegrationsbeziehungen zwischen trendbehafteten Zeitreihen werden häufig als ökonomische Gleichgewichtsbeziehungen interpretiert. Der Anpassungspfad zur langfristigen Gleichgewichtsbeziehung kann mit Hilfe eines Fehlerkorrekturmodells beschrieben werden. In der Veranstaltung sind praktische Übungen mit den Softwarepaketen Eviews und MulTi integriert. Univariate Analysis Description of time series Seasonal adjustment and trend filter ARMA-Models Nonstationary time series (ARIMA) Time varying variances (GARCH) Forecasting Multivariate Analysis VAR-Model Prediction and Causality Schätzung unter linearen Nebenbedingungen Lag order selection Structural (identified) VAR models Cointegration Spurious regression Asymptotic theory for nonstationary time series Two-step analysis of Engle-Granger Kointegrated VARs (Johansen’s ML analysis) Literatur Lütkepohl , H. (1991), Introduction to Multiple Time Series Analysis, Berlin: Springer Hamilton, J. (1994), Time Series Analysis, Princeton: Princeton University Press Lütkepohl , H. und M. Krätzig ( 2004 ), Applied Time Series Econometrics , Cambridge : Cambridge University Press.
Title: Generalized Non-Linear Strength Theory and Transformed Stress Space Author: Yao, Y. P.; Lu, D. C.; Zhou, A. N. ;Zou, B Source: Science in china series e-engineering materials science, 2004, 47(6): 691-709 Abstract: Based on the test data of frictional materials and previous research achievements in this field, a generalized non-linear strength theory (GNST) is proposed. It describes non-linear strength properties on the Pi-plane and the meridian plane using a unified formula, and it includes almost all the present non-linear strength theories, which can be used in just one material. The shape of failure function of the GNST is a smooth curve between the SMP criterion and the Mises criterion on the Pi-plane, and an exponential curve on the meridian plane. Through the transformed stress space based on the GNST, the combination of the GNST and various constitutive models using p and q as stress parameters can be realized simply and rationally in three-dimensional stress state.
最近跟风 Nobel Prize ,去 The Nobel Foundation 的官方网站 http://nobelprize.org 看了看,发现 Nobel 经济奖与时间序列分析有关系。 在 The Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel 2003 里 写着: http://nobelprize.org/nobel_prizes/economics/laureates/2003/index.html for methods of analyzing economic time series with time-varying volatility (ARCH) Robert F. Engle III ( November 10, 1942 ) for methods of analyzing economic time series with common trends (cointegration) Clive W.J. Granger (September 4, 1934 May 27, 2009) 这两位得主都是研究 analyzing economic time series 的。所以采用复杂时间序列分析新方法进行经济学研究(属于经济计量学 econometrics ),是有可能获得诺贝尔经济学奖。 Robert F. Engle III 研究的是:自回归条件异方差 (autoregressive conditional heteroskedasticity) 。 Clive W.J. Granger 研究的是:共整合 (cointegration) 。 他还是组合预测( The combination of forecasts )方法的主要提出者之一。 参考文献 : Robert F. Engle, http://weber.ucsd.edu/~mbacci/engle/index.html Clive W.J. Granger, http://www.econ.ucsd.edu/~cgranger/ The combination of forecasts, with J. Bates, Operational Research Quarterly , 20, 1969, 451-468. http://dss.ucsd.edu/~cgranger/pubs.html 欢迎补充! 欢迎纠正本博文的任何错误! 相关链接: 《什么样的人品得到诺贝尔经济奖》 http://www.sciencenet.cn/m/user_content.aspx?id=265938
参考文献: Microarray data analysis and mining approaches Francesca Cordero , Marco Botta and Raffaele A. Calogero Corresponding author: Raffaele A. Calogero, Department of Clinical and Biological Sciences, University of Torino, Italy. Briefings in Functional Genomics and Proteomics 2008 6(4):265-281; doi:10.1093/bfgp/elm034 Although two-sample differential expression analysis is probably the most common experiment, multi-series time-course microarray experiments are useful approaches for exploring biological processes. In these types of experiments, the researcher is frequently interested in studying gene expression changes over time and in evaluating trend differences between the various experimental groups . The large amount of data, multiplicity of experimental conditions and the dynamic nature of the experiments pose great challenges to data analysis. A comprehensive review of research in time series expression data analysis was published by Bar-Joseph in 2004 . Recently, maSigPro Conesa has published two methods for time-course microarray data analysis . One is maSigPro , and is part of Bioconductor packages. ...This will ultimately be used to find what are the profile differences between experimental groups. ANOVA-SCA The other is ANOVA-SCA and combines ANOVA-modeling and a dimension reduction technique to extract targeted signals from data bypassing structural noise. ANOVA-SCA basically applies PCA to the estimated parameters in each source of variation of an ANOVA model. ANOVASCA seems an effective approach for separating the data variability present in a complex time course experiment to extract the signal of interest from noisy data . fully Bayesian approach Angelini and coworkers have recently described a fully Bayesian approach to detect differentially expressed genes in time-course experiments. Their approach allows to explicitly use biological prior information and deals with various technical difficulties that arise in microarray timecourse experiments such as a small number of observations, non-uniform sampling intervals, missing or multiple data and temporal dependence between observations for each gene. Authors compared their method with that implemented in R-package time course and in the EDGE software claiming that their algorithm provides results which are much closer to a biologists choice and delivers a lower percentage of false positive and negative answers than other algorithms . Fischer and coworkers have compared methods for identifying differentially expressed genes on time-series microarray data simulated from artificial gene networks. They suggest the use of ANOVA variants of Cui and Churchill on the bases of simulated data and Efron and Tibshiranis empirical Bayes Wilcoxon rank sum test in the case experimental background cannot be effectively corrected. CHPM Shi has instead proposed an approach, based on a probabilistic continuous hidden process model ( CHPM ), to identify the various biological processes involved in a specific biological experiment. This method integrates time series expression data with GO biological processes , modelling the observed gene expression levels as being generated by a combination of multiple GO biological processes whose activity levels vary over time. OTHER TOOLs: STEM Short Time-series Expression Miner (STEM) (~8 time points or fewer). STEM allows researchers to identify significant temporal expression profiles and the genes associated with these profiles and to compare the behavior of these genes across multiple conditions. STEM is fully integrated with the Gene Ontology (GO) database supporting GO category gene enrichment analyses for sets of genes having the same temporal expression pattern. STEM also supports the ability to easily determine and visualize the behavior of genes belonging to a given GO category or user defined gene set, identifying which temporal expression profiles were enriched for these genes. DREM Dynamic Regulatory Events Miner (DREM) takes as input time series gene expression data and input that associates transcription factors with the genes they regulate . This regulatory information could come from Chromatin Immunoprecipitation (ChIP)-chip experiments or transcription factor binding site motif information. The transcription-factor gene regulation input does not need to be associated with specific time points. DREM after executing a computational method described in outputs an annotated dynamic regulatory map based on the data that can be interactively explored. The dynamic regulatory map highlights bifurcation events in the time series, that is places in the time series where sets of genes which previously had roughly similar expression level diverge. Often these bifurcation events can be explained by transcription factors selectively regulation a certain subset of genes. DREM annotates these events with transcription factors potentially responsible for them. DREM is related to another time series expression analysis software the Short Time-series Expression Miner ( STEM ) . While STEM focuses on identifying independent significant patterns in short time series data , DREM provides a global map of the gene regulation of the time series . The DREM method also incorporates in transcription factor-gene regulation information. DREM also is not limited to analyzing short time series data as STEM is. Some of the input file formats and options are the same in DREM as in STEM, in particular options related to gene filtering and the Gene Ontology analysis. For these aspects of DREM which are the same as in STEM, the relevant portions of the STEM manual have been incorporated into this manual. BiGGEsTS An integrated environment for the biclustering ( Madeira and Oliveira, 2004 ) analysis of time series gene expression data . Biclusters may be analyzed with Gene Ontology annotations to find out which contain statistically relevant biological information or even filtered or sorted according to several numerical and statistical criteria. 应用举例: STEM Comprehensive transcriptional profiling of NaCl-stressed Arabidopsis roots reveals novel classes of responsive genes Yuanqing Jiang and Michael K Deyholos* Address: Department of Biological Sciences, University of Alberta, Edmonton, Canada BMC Plant Biology 2006, 6:25 doi:10.1186/1471-2229-6-25 用STEM做了聚类,然后以GOTerm为层次标注各term中包含哪些cluster,有几个。 Combined analysis reveals a core set of cycling genes Yong Lu 1 , Shaun Mahony 2 , Panayiotis V Benos 2 , Roni Rosenfeld 3 , Itamar Simon 4 , Linda L Breeden 5 and Ziv Bar-Joseph 1 ,3 1 Department of Computer Science, Carnegie Mellon University, Forbes Avenue, Pittsburgh, Pennsylvania 15213, USA 2 Department of Computational Biology, University of Pittsburgh Medical School, Lothrop Street, Pittsburgh, Pennsylvania 15213, USA 3 Machine Learning Department, Carnegie Mellon University, Forbes Avenue, Pittsburgh, Pennsylvania 15213, USA 4 Department of Molecular Biology, Hebrew University Medical School, Jerusalem, Israel 91120 5 Basic Sciences Division, Fred Hutchinson Cancer Center, Fairview Avenue N, Seattle, Washington 98109, USA Genome Biology 2007, 8 : R146 doi:10.1186/gb-2007-8-7-r146 用STEM做了GO的显著性检验,P值,似乎大材小用 实战 STEM,总结如下: 1.用已经筛出来的差异基因做为输入,另外,用gene symbol来去除冗余(多个探针组对应同一蛋白,可以用UniRef);GO注释文件也以Gene symbol为ID(GO注释本来就是 注释 蛋白的,而非基因本身),但不可带有基因列表以外的Gene symbol。 2.可选STEM Cluster Method和K-means,设定Cluster数,进行GO Term富集分析。