<!DOCTYPE html> <html xmlns="http://www.w3.org/1999/xhtml"> <head> <meta charset="utf-8" /> <meta http-equiv="Content-Type" content="text/html; charset=utf-8" /> <meta name="generator" content="pandoc" /> <title>MASH v FLASH GTEx analysis: MASH v OHF</title> <script src="site_libs/jquery-1.11.3/jquery.min.js"></script> <meta name="viewport" content="width=device-width, initial-scale=1" /> <link href="site_libs/bootstrap-3.3.5/css/cosmo.min.css" rel="stylesheet" /> <script src="site_libs/bootstrap-3.3.5/js/bootstrap.min.js"></script> <script src="site_libs/bootstrap-3.3.5/shim/html5shiv.min.js"></script> <script src="site_libs/bootstrap-3.3.5/shim/respond.min.js"></script> <script src="site_libs/jqueryui-1.11.4/jquery-ui.min.js"></script> <link href="site_libs/tocify-1.9.1/jquery.tocify.css" rel="stylesheet" /> <script src="site_libs/tocify-1.9.1/jquery.tocify.js"></script> <script src="site_libs/navigation-1.1/tabsets.js"></script> <script src="site_libs/navigation-1.1/codefolding.js"></script> <link href="site_libs/highlightjs-9.12.0/textmate.css" rel="stylesheet" /> <script src="site_libs/highlightjs-9.12.0/highlight.js"></script> <link href="site_libs/font-awesome-4.5.0/css/font-awesome.min.css" rel="stylesheet" /> <style type="text/css">code{white-space: pre;}</style> <style type="text/css"> pre:not([class]) { background-color: white; } </style> <script type="text/javascript"> if (window.hljs) { hljs.configure({languages: []}); hljs.initHighlightingOnLoad(); if (document.readyState && document.readyState === "complete") { window.setTimeout(function() { hljs.initHighlighting(); }, 0); } } </script> <style type="text/css"> h1 { font-size: 34px; } h1.title { font-size: 38px; } h2 { font-size: 30px; } h3 { font-size: 24px; } h4 { font-size: 18px; } h5 { font-size: 16px; } h6 { font-size: 12px; } .table th:not([align]) { text-align: left; } </style> </head> <body> <style type = "text/css"> .main-container { max-width: 940px; margin-left: auto; margin-right: auto; } code { color: inherit; background-color: rgba(0, 0, 0, 0.04); } img { max-width:100%; height: auto; } .tabbed-pane { padding-top: 12px; } button.code-folding-btn:focus { outline: none; } </style> <style type="text/css"> /* padding for bootstrap navbar */ body { padding-top: 51px; padding-bottom: 40px; } /* offset scroll position for anchor links (for fixed navbar) */ .section h1 { padding-top: 56px; margin-top: -56px; } .section h2 { padding-top: 56px; margin-top: -56px; } .section h3 { padding-top: 56px; margin-top: -56px; } .section h4 { padding-top: 56px; margin-top: -56px; } .section h5 { padding-top: 56px; margin-top: -56px; } .section h6 { padding-top: 56px; margin-top: -56px; } </style> <script> // manage active state of menu based on current page $(document).ready(function () { // active menu anchor href = window.location.pathname href = href.substr(href.lastIndexOf('/') + 1) if (href === "") href = "index.html"; var menuAnchor = $('a[href="' + href + '"]'); // mark it active menuAnchor.parent().addClass('active'); // if it's got a parent navbar menu mark it active as well menuAnchor.closest('li.dropdown').addClass('active'); }); </script> <div class="container-fluid main-container"> <!-- tabsets --> <script> $(document).ready(function () { window.buildTabsets("TOC"); }); </script> <!-- code folding --> <style type="text/css"> .code-folding-btn { margin-bottom: 4px; } </style> <script> $(document).ready(function () { window.initializeCodeFolding("hide" === "show"); }); </script> <script> $(document).ready(function () { // move toc-ignore selectors from section div to header $('div.section.toc-ignore') .removeClass('toc-ignore') .children('h1,h2,h3,h4,h5').addClass('toc-ignore'); // establish options var options = { selectors: "h1,h2,h3", theme: "bootstrap3", context: '.toc-content', hashGenerator: function (text) { return text.replace(/[.\\/?&!#<>]/g, '').replace(/\s/g, '_').toLowerCase(); }, ignoreSelector: ".toc-ignore", scrollTo: 0 }; options.showAndHide = true; options.smoothScroll = true; // tocify var toc = $("#TOC").tocify(options).data("toc-tocify"); }); </script> <style type="text/css"> #TOC { margin: 25px 0px 20px 0px; } @media (max-width: 768px) { #TOC { position: relative; width: 100%; } } .toc-content { padding-left: 30px; padding-right: 40px; } div.main-container { max-width: 1200px; } div.tocify { width: 20%; max-width: 260px; max-height: 85%; } @media (min-width: 768px) and (max-width: 991px) { div.tocify { width: 25%; } } @media (max-width: 767px) { div.tocify { width: 100%; max-width: none; } } .tocify ul, .tocify li { line-height: 20px; } .tocify-subheader .tocify-item { font-size: 0.90em; padding-left: 25px; text-indent: 0; } .tocify .list-group-item { border-radius: 0px; } </style> <!-- setup 3col/9col grid for toc_float and main content --> <div class="row-fluid"> <div class="col-xs-12 col-sm-4 col-md-3"> <div id="TOC" class="tocify"> </div> </div> <div class="toc-content col-xs-12 col-sm-8 col-md-9"> <div class="navbar navbar-default navbar-fixed-top" role="navigation"> <div class="container"> <div class="navbar-header"> <button type="button" class="navbar-toggle collapsed" data-toggle="collapse" data-target="#navbar"> <span class="icon-bar"></span> <span class="icon-bar"></span> <span class="icon-bar"></span> </button> <a class="navbar-brand" href="index.html">MASHvFLASH</a> </div> <div id="navbar" class="navbar-collapse collapse"> <ul class="nav navbar-nav"> <li> <a href="index.html">Home</a> </li> <li> <a href="about.html">About</a> </li> </ul> <ul class="nav navbar-nav navbar-right"> <li> <a href="https://github.com/willwerscheid/MASHvFLASH"> <span class="fa fa-github"></span> </a> </li> </ul> </div><!--/.nav-collapse --> </div><!--/.container --> </div><!--/.navbar --> <!-- Add a small amount of space between sections. --> <style type="text/css"> div.section { padding-top: 12px; } </style> <div class="fluid-row" id="header"> <div class="btn-group pull-right"> <button type="button" class="btn btn-default btn-xs dropdown-toggle" data-toggle="dropdown" aria-haspopup="true" aria-expanded="false"><span>Code</span> <span class="caret"></span></button> <ul class="dropdown-menu" style="min-width: 50px;"> <li><a id="rmd-show-all-code" href="#">Show All Code</a></li> <li><a id="rmd-hide-all-code" href="#">Hide All Code</a></li> </ul> </div> <h1 class="title toc-ignore">MASH v FLASH GTEx analysis: MASH v OHF</h1> </div> <p><strong>Last updated:</strong> 2018-08-04</p> <strong>workflowr checks:</strong> <small>(Click a bullet for more information)</small> <ul> <li> <p><details> <summary> <strong style="color:blue;">✔</strong> <strong>R Markdown file:</strong> up-to-date </summary></p> <p>Great! Since the R Markdown file has been committed to the Git repository, you know the exact version of the code that produced these results.</p> </details> </li> <li> <p><details> <summary> <strong style="color:blue;">✔</strong> <strong>Environment:</strong> empty </summary></p> <p>Great job! The global environment was empty. Objects defined in the global environment can affect the analysis in your R Markdown file in unknown ways. For reproduciblity it’s best to always run the code in an empty environment.</p> </details> </li> <li> <p><details> <summary> <strong style="color:blue;">✔</strong> <strong>Seed:</strong> <code>set.seed(20180609)</code> </summary></p> <p>The command <code>set.seed(20180609)</code> was run prior to running the code in the R Markdown file. Setting a seed ensures that any results that rely on randomness, e.g. subsampling or permutations, are reproducible.</p> </details> </li> <li> <p><details> <summary> <strong style="color:blue;">✔</strong> <strong>Session information:</strong> recorded </summary></p> <p>Great job! Recording the operating system, R version, and package versions is critical for reproducibility.</p> </details> </li> <li> <p><details> <summary> <strong style="color:blue;">✔</strong> <strong>Repository version:</strong> <a href="https://github.com/willwerscheid/MASHvFLASH/tree/312476b990c8ec8b600434b175e84da2fb500c8b" target="_blank">312476b</a> </summary></p> Great! You are using Git for version control. Tracking code development and connecting the code version to the results is critical for reproducibility. The version displayed above was the version of the Git repository at the time these results were generated. <br><br> Note that you need to be careful to ensure that all relevant files for the analysis have been committed to Git prior to generating the results (you can use <code>wflow_publish</code> or <code>wflow_git_commit</code>). workflowr only checks the R Markdown file, but you know if there are other scripts or data files that it depends on. Below is the status of the Git repository when the results were generated: <pre><code> Ignored files: Ignored: .DS_Store Ignored: .Rhistory Ignored: .Rproj.user/ Ignored: data/ Ignored: docs/.DS_Store Ignored: docs/images/.DS_Store Ignored: docs/images/.Rapp.history Ignored: output/.DS_Store Ignored: output/.Rapp.history Ignored: output/MASHvFLASHgtex/.DS_Store Ignored: output/MASHvFLASHsims/.DS_Store Ignored: output/MASHvFLASHsims/backfit/.DS_Store Ignored: output/MASHvFLASHsims/backfit/.Rapp.history Untracked files: Untracked: docs/figure/OHFvTop20.Rmd/ Unstaged changes: Modified: code/gtexanalysis.R </code></pre> Note that any generated files, e.g. HTML, png, CSS, etc., are not included in this status report because it is ok for generated content to have uncommitted changes. </details> </li> </ul> <details> <summary> <small><strong>Expand here to see past versions:</strong></small> </summary> <ul> <table style="border-collapse:separate; border-spacing:5px;"> <thead> <tr> <th style="text-align:left;"> File </th> <th style="text-align:left;"> Version </th> <th style="text-align:left;"> Author </th> <th style="text-align:left;"> Date </th> <th style="text-align:left;"> Message </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> html </td> <td style="text-align:left;"> <a href="https://cdn.rawgit.com/willwerscheid/MASHvFLASH/5401954739b380856f398e937eb2939ee82e649e/docs/MASHvOHF.html" target="_blank">5401954</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-04 </td> <td style="text-align:left;"> Build site. </td> </tr> <tr> <td style="text-align:left;"> Rmd </td> <td style="text-align:left;"> <a href="https://github.com/willwerscheid/MASHvFLASH/blob/b2aa853ac2739a95bc559d79edf918167c97833d/analysis/MASHvOHF.Rmd" target="_blank">b2aa853</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-04 </td> <td style="text-align:left;"> wflow_publish(c(“analysis/MASHvTop20.Rmd”, </td> </tr> <tr> <td style="text-align:left;"> html </td> <td style="text-align:left;"> <a href="https://cdn.rawgit.com/willwerscheid/MASHvFLASH/ea857fed6060fadfe06b9b6f4d7273ec8b960823/docs/MASHvOHF.html" target="_blank">ea857fe</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-03 </td> <td style="text-align:left;"> Build site. </td> </tr> <tr> <td style="text-align:left;"> Rmd </td> <td style="text-align:left;"> <a href="https://github.com/willwerscheid/MASHvFLASH/blob/c5d0c86fb2b720f9eb6b652eec9b6428cebecfc2/analysis/MASHvOHF.Rmd" target="_blank">c5d0c86</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-03 </td> <td style="text-align:left;"> wflow_publish(“analysis/MASHvOHF.Rmd”) </td> </tr> <tr> <td style="text-align:left;"> html </td> <td style="text-align:left;"> <a href="https://cdn.rawgit.com/willwerscheid/MASHvFLASH/e78875ccee6156479deb7b221b05bf43f1e9d442/docs/MASHvOHF.html" target="_blank">e78875c</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-02 </td> <td style="text-align:left;"> Build site. </td> </tr> <tr> <td style="text-align:left;"> Rmd </td> <td style="text-align:left;"> <a href="https://github.com/willwerscheid/MASHvFLASH/blob/0e165ca9657c7b0798613ad8914cd2b5cc442da7/analysis/MASHvOHF.Rmd" target="_blank">0e165ca</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-02 </td> <td style="text-align:left;"> wflow_publish(c(“analysis/MASHvOHF.Rmd”, “analysis/index.Rmd”)) </td> </tr> <tr> <td style="text-align:left;"> html </td> <td style="text-align:left;"> <a href="https://cdn.rawgit.com/willwerscheid/MASHvFLASH/a91ab89034a324bfb9412b3142d4418ee3064b30/docs/MASHvOHF.html" target="_blank">a91ab89</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-02 </td> <td style="text-align:left;"> Build site. </td> </tr> <tr> <td style="text-align:left;"> Rmd </td> <td style="text-align:left;"> <a href="https://github.com/willwerscheid/MASHvFLASH/blob/cebcb57aa34794f6e25155d34f70d9585b2dafa6/analysis/MASHvOHF.Rmd" target="_blank">cebcb57</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-02 </td> <td style="text-align:left;"> wflow_publish(“analysis/MASHvOHF.Rmd”) </td> </tr> </tbody> </table> </ul> <p></details></p> <hr /> <div id="introduction" class="section level2"> <h2>Introduction</h2> <p>In the <a href="MASHvFLASHgtex2.html">previous analysis</a>, I proposed several workflows for fitting MASH and FLASH objects to GTEx data. Here and in the next few analyses I examine differences among fits. I begin by comparing the MASH fit to the OHF fit.</p> <p>For the code used to obtain the fits, see the <a href="MASHvFLASHgtex2.html#code">previous analysis</a>.</p> <pre class="r"><code>library(mashr)</code></pre> <pre><code>Loading required package: ashr</code></pre> <pre class="r"><code>devtools::load_all("/Users/willwerscheid/GitHub/flashr/")</code></pre> <pre><code>Loading flashr</code></pre> <pre class="r"><code>gtex <- readRDS(gzcon(url("https://github.com/stephenslab/gtexresults/blob/master/data/MatrixEQTLSumStats.Portable.Z.rds?raw=TRUE"))) strong <- t(gtex$strong.z) fpath <- "./output/MASHvFLASHgtex2/" m_final <- readRDS(paste0(fpath, "m.rds")) fl_final <- readRDS(paste0(fpath, "OHF.rds")) m_lfsr <- t(get_lfsr(m_final)) m_pm <- t(get_pm(m_final)) fl_lfsr <- readRDS(paste0(fpath, "fllfsr.rds")) fl_lfsr <- fl_lfsr[[1]] fl_pm <- flash_get_fitted_values(fl_final)</code></pre> </div> <div id="introduction-to-plots" class="section level2"> <h2>Introduction to plots</h2> <p>The main tool I will use in these analyses is a function that plots observed values vs. MASH or FLASH posterior means for a single test (over all 44 conditions).</p> <p>The observed <span class="math inline">\(z\)</span>-scores are plotted as hollow circles. Posterior means are plotted as triangles if the effect is judged to be significant (LFSR <span class="math inline">\(\le .05\)</span>) and as squares otherwise. Larger triangles indicate lower LFSRs, with the largest triangles corresponding to “highly significant” effects (LFSR <span class="math inline">\(\le .01\)</span>). The posterior means are colored using the GTEx colors used in previous analyses.</p> <pre class="r"><code>missing.tissues <- c(7, 8, 19, 20, 24, 25, 31, 34, 37) gtex.colors <- read.table("https://github.com/stephenslab/gtexresults/blob/master/data/GTExColors.txt?raw=TRUE", sep = '\t', comment.char = '')[-missing.tissues, 2] OHF.colors <- c("tan4", "tan3") zero.colors <- c("black", gray.colors(19, 0.2, 0.9), gray.colors(17, 0.95, 1)) plot_test <- function(n, lfsr, pm, method_name) { plot(strong[, n], pch=1, col="black", xlab="", ylab="", cex=0.6, ylim=c(min(c(strong[, n], 0)), max(c(strong[, n], 0))), main=paste0("Test #", n, "; ", method_name)) size = rep(0.6, 44) shape = rep(15, 44) signif <- lfsr[, n] <= .05 shape[signif] <- 17 size[signif] <- 1.35 - 15 * lfsr[signif, n] size <- pmin(size, 1.2) points(pm[, n], pch=shape, col=as.character(gtex.colors), cex=size) abline(0, 0) } plot_ohf_v_ohl_loadings <- function(n, ohf_fit, ohl_fit, ohl_name, legend_pos = "bottomright") { ohf <- abs(ohf_fit$EF[n, ] * apply(abs(ohf_fit$EL), 2, max)) ohl <- -abs(ohl_fit$EF[n, ] * apply(abs(ohl_fit$EL), 2, max)) data <- rbind(c(ohf, rep(0, length(ohl) - 45)), c(ohl[1:45], rep(0, length(ohf) - 45), ohl[46:length(ohl)])) colors <- c("black", as.character(gtex.colors), OHF.colors, zero.colors[1:(length(ohl) - 45)]) x <- barplot(data, beside=T, col=rep(colors, each=2), main=paste0("Test #", n, " loadings"), legend.text = c("OHF", ohl_name), args.legend = list(x = legend_pos, bty = "n", pch="+-", fill=NULL, border="white")) text(x[2*(46:47) - 1], min(data) / 10, labels=as.character(1:2), cex=0.4) text(x[2*(48:ncol(data))], max(data) / 10, labels=as.character(1:(length(ohl) - 45)), cex=0.4) } compare_methods <- function(lfsr1, lfsr2, pm1, pm2) { res <- list() res$first_not_second <- find_A_not_B(lfsr1, lfsr2) res$lg_first_not_second <- find_large_A_not_B(lfsr1, lfsr2) res$second_not_first <- find_A_not_B(lfsr2, lfsr1) res$lg_second_not_first <- find_large_A_not_B(lfsr2, lfsr1) res$diff_pms <- find_overall_pm_diff(pm1, pm2) return(res) } # Find tests where many conditions are significant according to # method A but not according to method B. find_A_not_B <- function(lfsrA, lfsrB) { select_tests(colSums(lfsrA <= 0.05 & lfsrB > 0.05)) } # Find tests where many conditions are highly significant according to # method A but are not significant according to method B. find_large_A_not_B <- function(lfsrA, lfsrB) { select_tests(colSums(lfsrA <= 0.01 & lfsrB > 0.05)) } find_overall_pm_diff <- function(pmA, pmB, n = 4) { pm_diff <- colSums((pmA - pmB)^2) return(order(pm_diff, decreasing = TRUE)[1:4]) } # Get at least four (or min_n) "top" tests. select_tests <- function(colsums, min_n = 4) { n <- 45 n_tests <- 0 while (n_tests < min_n && n > 0) { n <- n - 1 n_tests <- sum(colsums >= n) } return(which(colsums >= n)) }</code></pre> </div> <div id="significant-for-ohf-not-mash" class="section level2"> <h2>Significant for OHF, not MASH</h2> <p>By far the most common case is one in which FLASH finds a combination of a small equal effect and a large unique effect, while MASH only finds the unique effect to be significant. (Here, it is possibly relevant to recall that my <a href="MASHvFLASHsims2.html">simulation study</a> suggested that FLASH outperforms MASH when the “true” effect is a combination of an equal effect and a unique effect.) In such cases, FLASH usually (but not always) applies far less shrinkage to the unique effect. Some typical examples follow.</p> <pre class="r"><code># interesting.tests <- compare_methods(fl_lfsr, m_lfsr, fl_pm, m_pm) par(mfrow=c(1, 2)) identical.plus.unique <- c(15828, 1480, 15711, 10000) for (n in identical.plus.unique) { plot_test(n, fl_lfsr, fl_pm, "OHF") plot_test(n, m_lfsr, m_pm, "MASH") }</code></pre> <p><img src="figure/MASHvOHF.Rmd/ohf_not_mash-1.png" width="672" style="display: block; margin: auto;" /></p> <details> <summary><em>Expand here to see past versions of ohf_not_mash-1.png:</em></summary> <table style="border-collapse:separate; border-spacing:5px;"> <thead> <tr> <th style="text-align:left;"> Version </th> <th style="text-align:left;"> Author </th> <th style="text-align:left;"> Date </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> <a href="https://github.com/willwerscheid/MASHvFLASH/blob/a91ab89034a324bfb9412b3142d4418ee3064b30/docs/figure/MASHvOHF.Rmd/ohf_not_mash-1.png" target="_blank">a91ab89</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-02 </td> </tr> </tbody> </table> <p></details> <img src="figure/MASHvOHF.Rmd/ohf_not_mash-2.png" width="672" style="display: block; margin: auto;" /></p> <details> <summary><em>Expand here to see past versions of ohf_not_mash-2.png:</em></summary> <table style="border-collapse:separate; border-spacing:5px;"> <thead> <tr> <th style="text-align:left;"> Version </th> <th style="text-align:left;"> Author </th> <th style="text-align:left;"> Date </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> <a href="https://github.com/willwerscheid/MASHvFLASH/blob/a91ab89034a324bfb9412b3142d4418ee3064b30/docs/figure/MASHvOHF.Rmd/ohf_not_mash-2.png" target="_blank">a91ab89</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-02 </td> </tr> </tbody> </table> <p></details> <img src="figure/MASHvOHF.Rmd/ohf_not_mash-3.png" width="672" style="display: block; margin: auto;" /></p> <details> <summary><em>Expand here to see past versions of ohf_not_mash-3.png:</em></summary> <table style="border-collapse:separate; border-spacing:5px;"> <thead> <tr> <th style="text-align:left;"> Version </th> <th style="text-align:left;"> Author </th> <th style="text-align:left;"> Date </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> <a href="https://github.com/willwerscheid/MASHvFLASH/blob/a91ab89034a324bfb9412b3142d4418ee3064b30/docs/figure/MASHvOHF.Rmd/ohf_not_mash-3.png" target="_blank">a91ab89</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-02 </td> </tr> </tbody> </table> <p></details> <img src="figure/MASHvOHF.Rmd/ohf_not_mash-4.png" width="672" style="display: block; margin: auto;" /></p> <details> <summary><em>Expand here to see past versions of ohf_not_mash-4.png:</em></summary> <table style="border-collapse:separate; border-spacing:5px;"> <thead> <tr> <th style="text-align:left;"> Version </th> <th style="text-align:left;"> Author </th> <th style="text-align:left;"> Date </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> <a href="https://github.com/willwerscheid/MASHvFLASH/blob/a91ab89034a324bfb9412b3142d4418ee3064b30/docs/figure/MASHvOHF.Rmd/ohf_not_mash-4.png" target="_blank">a91ab89</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-02 </td> </tr> </tbody> </table> <p></details></p> </div> <div id="significant-for-mash-not-ohf" class="section level2"> <h2>Significant for MASH, not OHF</h2> <p>To understand the typical situation where MASH declares effects to be significant but OHF does not, recall from my <a href="MASHvFLASHgtex2.html">analysis of GTEx workflows</a> that the MASH fit puts large mixture weights on data-driven covariance matrices (around 0.6 on “ED_tPCA” and 0.25 on “ED_PCA_2”) and a comparatively small weight (around 0.1) on the “equal effects” covariance structure. In contrast, the equal effects loading accounts for about 70% of the variance explained by the OHF fit. This, I think, is why OHF is so much more likely to find an equal effect in the tests above. MASH, on the contrary, is more likely to find a pattern of covariance that derives from the data-driven structures. Notice the similarity of the MASH estimates in the following tests (in the last example, signs are reversed):</p> <pre class="r"><code>par(mfrow=c(1, 2)) mash.covar <- c(2868, 9716, 10368, 9716) for (n in mash.covar) { plot_test(n, fl_lfsr, fl_pm, "OHF") plot_test(n, m_lfsr, m_pm, "MASH") }</code></pre> <p><img src="figure/MASHvOHF.Rmd/mash_not_ohf-1.png" width="672" style="display: block; margin: auto;" /></p> <details> <summary><em>Expand here to see past versions of mash_not_ohf-1.png:</em></summary> <table style="border-collapse:separate; border-spacing:5px;"> <thead> <tr> <th style="text-align:left;"> Version </th> <th style="text-align:left;"> Author </th> <th style="text-align:left;"> Date </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> <a href="https://github.com/willwerscheid/MASHvFLASH/blob/a91ab89034a324bfb9412b3142d4418ee3064b30/docs/figure/MASHvOHF.Rmd/mash_not_ohf-1.png" target="_blank">a91ab89</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-02 </td> </tr> </tbody> </table> <p></details> <img src="figure/MASHvOHF.Rmd/mash_not_ohf-2.png" width="672" style="display: block; margin: auto;" /></p> <details> <summary><em>Expand here to see past versions of mash_not_ohf-2.png:</em></summary> <table style="border-collapse:separate; border-spacing:5px;"> <thead> <tr> <th style="text-align:left;"> Version </th> <th style="text-align:left;"> Author </th> <th style="text-align:left;"> Date </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> <a href="https://github.com/willwerscheid/MASHvFLASH/blob/a91ab89034a324bfb9412b3142d4418ee3064b30/docs/figure/MASHvOHF.Rmd/mash_not_ohf-2.png" target="_blank">a91ab89</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-02 </td> </tr> </tbody> </table> <p></details> <img src="figure/MASHvOHF.Rmd/mash_not_ohf-3.png" width="672" style="display: block; margin: auto;" /></p> <details> <summary><em>Expand here to see past versions of mash_not_ohf-3.png:</em></summary> <table style="border-collapse:separate; border-spacing:5px;"> <thead> <tr> <th style="text-align:left;"> Version </th> <th style="text-align:left;"> Author </th> <th style="text-align:left;"> Date </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> <a href="https://github.com/willwerscheid/MASHvFLASH/blob/a91ab89034a324bfb9412b3142d4418ee3064b30/docs/figure/MASHvOHF.Rmd/mash_not_ohf-3.png" target="_blank">a91ab89</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-02 </td> </tr> </tbody> </table> <p></details> <img src="figure/MASHvOHF.Rmd/mash_not_ohf-4.png" width="672" style="display: block; margin: auto;" /></p> <details> <summary><em>Expand here to see past versions of mash_not_ohf-4.png:</em></summary> <table style="border-collapse:separate; border-spacing:5px;"> <thead> <tr> <th style="text-align:left;"> Version </th> <th style="text-align:left;"> Author </th> <th style="text-align:left;"> Date </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> <a href="https://github.com/willwerscheid/MASHvFLASH/blob/a91ab89034a324bfb9412b3142d4418ee3064b30/docs/figure/MASHvOHF.Rmd/mash_not_ohf-4.png" target="_blank">a91ab89</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-02 </td> </tr> </tbody> </table> <p></details></p> <p>At first, it seems bizarre that effects with posterior means near zero are judged to be significant. My guess is that in each of these cases, the observations follow a pattern that closely matches the data-driven covariance structures, so that MASH can be highly confident in the sign of the effect even if the effect is not very large.</p> </div> <div id="different-posterior-means" class="section level2"> <h2>Different posterior means</h2> <p>I conclude with some examples of tests with a large mean squared difference in posterior means. As might be expected, these are all tests where effect sizes are large for many conditions. Interestingly, MASH applies little to no shrinkage in such cases (even for small effects). FLASH does a better job applying a reasonable amount of shrinkage to small and moderate effects, but can be overly aggressive in shrinking large effects.</p> <p>In test #2857, for example, FLASH applies a huge amount of shrinkage to the large effects observed in visceral adipose omental tissue and in mammary tissue, shrinking the <span class="math inline">\(z\)</span>-scores from -9.0 to -2.3 and from -8.1 to -2.3, respectively. These are both tissues for which the prior on the unique effect is heavily concentrated near zero (see <a href="MASHvFLASHgtex2.html#priors_on_factors_(steps_2-3)">here</a>), so the results make sense given the FLASH fit, but the shrinkage still seems extreme. I think that this problem could be mitigated by training on a larger subset of random tests (so that some examples of large effects in these tissues get included).</p> <p>In contrast, the MASH posterior mixture weights for test #2857 are heavily concentrated on the <code>simple_het</code> matrices, which posit only weak correlations among conditions.</p> <pre class="r"><code>par(mfrow=c(1, 2)) diff.pms <- c(2857, 697, 384, 1680) for (n in diff.pms) { plot_test(n, fl_lfsr, fl_pm, "OHF") plot_test(n, m_lfsr, m_pm, "MASH") }</code></pre> <p><img src="figure/MASHvOHF.Rmd/diff_pms-1.png" width="672" style="display: block; margin: auto;" /></p> <details> <summary><em>Expand here to see past versions of diff_pms-1.png:</em></summary> <table style="border-collapse:separate; border-spacing:5px;"> <thead> <tr> <th style="text-align:left;"> Version </th> <th style="text-align:left;"> Author </th> <th style="text-align:left;"> Date </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> <a href="https://github.com/willwerscheid/MASHvFLASH/blob/ea857fed6060fadfe06b9b6f4d7273ec8b960823/docs/figure/MASHvOHF.Rmd/diff_pms-1.png" target="_blank">ea857fe</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-03 </td> </tr> </tbody> </table> <p></details> <img src="figure/MASHvOHF.Rmd/diff_pms-2.png" width="672" style="display: block; margin: auto;" /></p> <details> <summary><em>Expand here to see past versions of diff_pms-2.png:</em></summary> <table style="border-collapse:separate; border-spacing:5px;"> <thead> <tr> <th style="text-align:left;"> Version </th> <th style="text-align:left;"> Author </th> <th style="text-align:left;"> Date </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> <a href="https://github.com/willwerscheid/MASHvFLASH/blob/ea857fed6060fadfe06b9b6f4d7273ec8b960823/docs/figure/MASHvOHF.Rmd/diff_pms-2.png" target="_blank">ea857fe</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-03 </td> </tr> </tbody> </table> <p></details> <img src="figure/MASHvOHF.Rmd/diff_pms-3.png" width="672" style="display: block; margin: auto;" /></p> <details> <summary><em>Expand here to see past versions of diff_pms-3.png:</em></summary> <table style="border-collapse:separate; border-spacing:5px;"> <thead> <tr> <th style="text-align:left;"> Version </th> <th style="text-align:left;"> Author </th> <th style="text-align:left;"> Date </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> <a href="https://github.com/willwerscheid/MASHvFLASH/blob/ea857fed6060fadfe06b9b6f4d7273ec8b960823/docs/figure/MASHvOHF.Rmd/diff_pms-3.png" target="_blank">ea857fe</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-03 </td> </tr> </tbody> </table> <p></details> <img src="figure/MASHvOHF.Rmd/diff_pms-4.png" width="672" style="display: block; margin: auto;" /></p> <details> <summary><em>Expand here to see past versions of diff_pms-4.png:</em></summary> <table style="border-collapse:separate; border-spacing:5px;"> <thead> <tr> <th style="text-align:left;"> Version </th> <th style="text-align:left;"> Author </th> <th style="text-align:left;"> Date </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> <a href="https://github.com/willwerscheid/MASHvFLASH/blob/ea857fed6060fadfe06b9b6f4d7273ec8b960823/docs/figure/MASHvOHF.Rmd/diff_pms-4.png" target="_blank">ea857fe</a> </td> <td style="text-align:left;"> Jason Willwerscheid </td> <td style="text-align:left;"> 2018-08-03 </td> </tr> </tbody> </table> <p></details></p> </div> <div id="session-information" class="section level2"> <h2>Session information</h2> <pre class="r"><code>sessionInfo()</code></pre> <pre><code>R version 3.4.3 (2017-11-30) Platform: x86_64-apple-darwin15.6.0 (64-bit) Running under: macOS High Sierra 10.13.6 Matrix products: default BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib locale: [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8 attached base packages: [1] stats graphics grDevices utils datasets methods base other attached packages: [1] flashr_0.5-12 mashr_0.2-7 ashr_2.2-10 loaded via a namespace (and not attached): [1] Rcpp_0.12.17 pillar_1.2.1 compiler_3.4.3 [4] git2r_0.21.0 plyr_1.8.4 workflowr_1.0.1 [7] R.methodsS3_1.7.1 R.utils_2.6.0 iterators_1.0.9 [10] tools_3.4.3 testthat_2.0.0 digest_0.6.15 [13] tibble_1.4.2 gtable_0.2.0 evaluate_0.10.1 [16] memoise_1.1.0 lattice_0.20-35 rlang_0.2.0 [19] Matrix_1.2-12 foreach_1.4.4 commonmark_1.4 [22] yaml_2.1.17 parallel_3.4.3 ebnm_0.1-12 [25] mvtnorm_1.0-7 xml2_1.2.0 withr_2.1.1.9000 [28] stringr_1.3.0 knitr_1.20 roxygen2_6.0.1.9000 [31] devtools_1.13.4 rprojroot_1.3-2 grid_3.4.3 [34] R6_2.2.2 rmarkdown_1.8 rmeta_3.0 [37] ggplot2_2.2.1 magrittr_1.5 whisker_0.3-2 [40] scales_0.5.0 backports_1.1.2 codetools_0.2-15 [43] htmltools_0.3.6 MASS_7.3-48 assertthat_0.2.0 [46] softImpute_1.4 colorspace_1.3-2 stringi_1.1.6 [49] lazyeval_0.2.1 munsell_0.4.3 doParallel_1.0.11 [52] pscl_1.5.2 truncnorm_1.0-8 SQUAREM_2017.10-1 [55] R.oo_1.21.0 </code></pre> </div> <!-- Adjust MathJax settings so that all math formulae are shown using TeX fonts only; see http://docs.mathjax.org/en/latest/configuration.html. This will make the presentation more consistent at the cost of the webpage sometimes taking slightly longer to load. Note that this only works because the footer is added to webpages before the MathJax javascript. --> <script type="text/x-mathjax-config"> MathJax.Hub.Config({ "HTML-CSS": { availableFonts: ["TeX"] } }); </script> <hr> <p> This reproducible <a href="http://rmarkdown.rstudio.com">R Markdown</a> analysis was created with <a href="https://github.com/jdblischak/workflowr">workflowr</a> 1.0.1 </p> <hr> </div> </div> </div> <script> // add bootstrap table styles to pandoc tables function bootstrapStylePandocTables() { $('tr.header').parent('thead').parent('table').addClass('table table-condensed'); } $(document).ready(function () { bootstrapStylePandocTables(); }); </script> <!-- dynamically load mathjax for compatibility with self-contained --> <script> (function () { var script = document.createElement("script"); script.type = "text/javascript"; script.src = "https://mathjax.rstudio.com/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"; document.getElementsByTagName("head")[0].appendChild(script); })(); </script> </body> </html>