<!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" />


<meta name="author" content="Jason Willwerscheid" />


<title>Mash vs. Flash</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>
<li>
  <a href="license.html">License</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 vs. Flash</h1>
<h4 class="author"><em>Jason Willwerscheid</em></h4>
<h4 class="date"><em>4/20/2018</em></h4>

</div>


<p><strong>Last updated:</strong> 2018-06-11</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/485717cda6cfd74224352868188e136762fc2855" target="_blank">485717c</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:    .Rproj.user/

</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;">
Rmd
</td>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/485717cda6cfd74224352868188e136762fc2855/analysis/mashvflash.Rmd" target="_blank">485717c</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</td>
<td style="text-align:left;">
wflow_publish(“analysis/mashvflash.Rmd”)
</td>
</tr>
<tr>
<td style="text-align:left;">
html
</td>
<td style="text-align:left;">
<a href="https://cdn.rawgit.com/willwerscheid/MASHvFLASH/a1180b51f9cae5996e43a7bb368317bad76cf981/docs/mashvflash.html" target="_blank">a1180b5</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</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/76285202554d56fc0ea43e593608e626dbdbf364/analysis/mashvflash.Rmd" target="_blank">7628520</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</td>
<td style="text-align:left;">
wflow_publish(“analysis/mashvflash.Rmd”)
</td>
</tr>
<tr>
<td style="text-align:left;">
html
</td>
<td style="text-align:left;">
<a href="https://cdn.rawgit.com/willwerscheid/MASHvFLASH/624de1dfb3db11bdc40a4a3d946a8fbd7abf40ce/docs/mashvflash.html" target="_blank">624de1d</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</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/ea2721a4c6fa3ab74883bf94c6c6687d0457efb9/analysis/mashvflash.Rmd" target="_blank">ea2721a</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</td>
<td style="text-align:left;">
wflow_publish(“analysis/mashvflash.Rmd”)
</td>
</tr>
<tr>
<td style="text-align:left;">
html
</td>
<td style="text-align:left;">
<a href="https://cdn.rawgit.com/willwerscheid/MASHvFLASH/9f81f0649020df46c440e91d61356176321aa930/docs/mashvflash.html" target="_blank">9f81f06</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</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/f5750b8f736c2bede897df1fe7ce9a82db1e1552/analysis/mashvflash.Rmd" target="_blank">f5750b8</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</td>
<td style="text-align:left;">
wflow_publish(“analysis/mashvflash.Rmd”)
</td>
</tr>
<tr>
<td style="text-align:left;">
html
</td>
<td style="text-align:left;">
<a href="https://cdn.rawgit.com/willwerscheid/MASHvFLASH/17f4dfb20b57e7e3b41a94a7b2dce787af325b3a/docs/mashvflash.html" target="_blank">17f4dfb</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</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/b47adc3c271eb4265a971fbc02529baa9a428286/analysis/mashvflash.Rmd" target="_blank">b47adc3</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</td>
<td style="text-align:left;">
wflow_publish(“analysis/mashvflash.Rmd”)
</td>
</tr>
<tr>
<td style="text-align:left;">
html
</td>
<td style="text-align:left;">
<a href="https://cdn.rawgit.com/willwerscheid/MASHvFLASH/6508f2be3666427c3f8bd6c348e59ab7b9355565/docs/mashvflash.html" target="_blank">6508f2b</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-09
</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/861d4b3595125a0b8a304aa13375835acf4c8505/analysis/mashvflash.Rmd" target="_blank">861d4b3</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-09
</td>
<td style="text-align:left;">
readd mashvflash analysis
</td>
</tr>
<tr>
<td style="text-align:left;">
html
</td>
<td style="text-align:left;">
<a href="https://cdn.rawgit.com/willwerscheid/MASHvFLASH/861d4b3595125a0b8a304aa13375835acf4c8505/docs/mashvflash.html" target="_blank">861d4b3</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-09
</td>
<td style="text-align:left;">
readd mashvflash analysis
</td>
</tr>
<tr>
<td style="text-align:left;">
Rmd
</td>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/a76f3cf9f3dad726abb36bc409df06e9c31f57d2/analysis/mashvflash.Rmd" target="_blank">a76f3cf</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-09
</td>
<td style="text-align:left;">
add mashvflash rmd
</td>
</tr>
<tr>
<td style="text-align:left;">
html
</td>
<td style="text-align:left;">
<a href="https://cdn.rawgit.com/willwerscheid/MASHvFLASH/a76f3cf9f3dad726abb36bc409df06e9c31f57d2/docs/mashvflash.html" target="_blank">a76f3cf</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-09
</td>
<td style="text-align:left;">
add mashvflash rmd
</td>
</tr>
</tbody>
</table>
</ul>
<p></details></p>
<hr />
<div id="code" class="section level2">
<h2>Code</h2>
<p>Setup code.</p>
<pre class="r"><code># FLASH v MASH ------------------------------------------------------
flash_v_mash &lt;- function(Y, true_Y, nfactors) {
  data &lt;- flash_set_data(Y, S = 1)
  res &lt;- list()

  t &lt;- Sys.time()
  fl &lt;- fit_flash(data, nfactors)
  res$fl_time &lt;- Sys.time() - t

  t &lt;- Sys.time()
  m &lt;- fit_mash(Y)
  res$m_time &lt;- Sys.time() - t

  # Sample from FLASH fit
  fl_sampler &lt;- flash_lf_sampler(Y, fl, ebnm_fn=ebnm_pn, fixed=&quot;factors&quot;)

  nsamp &lt;- 200
  fl_samp &lt;- fl_sampler(nsamp)

  res$fl_mse &lt;- flash_pm_mse(fl_samp, true_Y)
  res$m_mse &lt;- mash_pm_mse(m, true_Y)
  res$fl_ci &lt;- flash_ci_acc(fl_samp, true_Y)
  res$m_ci &lt;- mash_ci_acc(m, true_Y)
  res$fl_lfsr &lt;- flash_lfsr(fl_samp, true_Y)
  res$m_lfsr &lt;-  mash_lfsr(m, true_Y)
  res
}

plot_res &lt;- function(res) {
  old_par &lt;- par(&quot;mfrow&quot;)
  par(mfrow=c(1, 2))
  x &lt;- seq(0.025, 0.475, by=0.05)
  plot(x, res$fl_lfsr, type=&#39;l&#39;, ylim=c(0, 0.6), xlab=&quot;FLASH&quot;, ylab=&quot;lfsr&quot;)
  abline(0, 1)
  plot(x, res$m_lfsr, type=&#39;l&#39;, ylim=c(0, 0.6), xlab=&quot;MASH&quot;, ylab=&quot;lfsr&quot;)
  abline(0, 1)
  par(mfrow=old_par)
}


# Fit using FLASH ---------------------------------------------------
fit_flash &lt;- function(data, nfactors) {
  p &lt;- ncol(data$Y)
  fl &lt;- flash_add_greedy(data, nfactors, var_type = &quot;zero&quot;)
  fl &lt;- flash_add_fixed_f(data, diag(rep(1, p)), fl)
  flash_backfit(data, fl, nullcheck = F, var_type = &quot;zero&quot;)
}

# Fit using MASH ---------------------------------------------------
fit_mash &lt;- function(Y) {
  data &lt;- mash_set_data(Y)
  U.c = cov_canonical(data)
  m.1by1 &lt;- mash_1by1(data)
  strong &lt;- get_significant_results(m.1by1, 0.05)
  U.pca &lt;- cov_pca(data, 5, strong)
  U.ed &lt;- cov_ed(data, U.pca, strong)
  mash(data, c(U.c,U.ed))
}


# MSE of posterior means (FLASH) ------------------------------------
flash_pm_mse &lt;- function(fl_samp, true_Y) {
  n &lt;- nrow(true_Y)
  p &lt;- ncol(true_Y)
  nsamp &lt;- length(fl_samp)

  post_means &lt;- matrix(0, nrow=n, ncol=p)
  for (i in 1:nsamp) {
    post_means &lt;- post_means + fl_samp[[i]]
  }
  post_means &lt;- post_means / nsamp
  sum((post_means - true_Y)^2) / (n * p)
}
# Compare with just using FLASH LF:
# sum((flash_get_lf(fl)- true_flash_Y)^2) / (n * p)


# MSE for MASH ------------------------------------------------------
mash_pm_mse &lt;- function(m, true_Y) {
  n &lt;- nrow(true_Y)
  p &lt;- ncol(true_Y)
  sum((get_pm(m) - true_Y)^2) / (n * p)
}


# CI coverage for FLASH ---------------------------------------------
flash_ci_acc &lt;- function(fl_samp, true_Y) {
  n &lt;- nrow(true_Y)
  p &lt;- ncol(true_Y)
  nsamp &lt;- length(fl_samp)

  flat_samp &lt;- matrix(0, nrow=n*p, ncol=nsamp)
  for (i in 1:nsamp) {
    flat_samp[, i] &lt;- as.vector(fl_samp[[i]])
  }
  CI &lt;- t(apply(flat_samp, 1, function(x) {quantile(x, c(0.025, 0.975))}))
  sum((as.vector(true_Y) &gt; CI[, 1])
      &amp; (as.vector(true_Y &lt; CI[, 2]))) / (n * p)
}

# CI coverage for MASH ----------------------------------------------
mash_ci_acc &lt;- function(m, true_Y) {
  sum((true_Y &gt; get_pm(m) - 1.96 * get_psd(m))
      &amp; (true_Y &lt; get_pm(m) + 1.96 * get_psd(m))) / (n * p)
}


# LFSR for FLASH ----------------------------------------------------
flash_lfsr &lt;- function(fl_samp, true_Y, step=0.05) {
  n &lt;- nrow(true_Y)
  p &lt;- ncol(true_Y)
  nsamp &lt;- length(fl_samp)

  lfsr &lt;- matrix(0, nrow=n, ncol=p)
  for (i in 1:nsamp) {
    lfsr &lt;- lfsr + (fl_samp[[i]] &gt; 0) + 0.5*(fl_samp[[i]] == 0)
  }
  signs &lt;- lfsr &gt;= nsamp / 2
  correct_signs &lt;- true_Y &gt; 0
  gotitright &lt;- signs == correct_signs
  lfsr &lt;- pmin(lfsr, 100 - lfsr) / 100

  nsteps &lt;- floor(.5 / step)
  fsr_by_lfsr &lt;- rep(0, nsteps)
  for (k in 1:nsteps) {
    idx &lt;- (lfsr &gt;= (step * (k - 1)) &amp; lfsr &lt; (step * k))
    fsr_by_lfsr[k] &lt;- ifelse(sum(idx) == 0, 0,
                             1 - sum(gotitright[idx]) / sum(idx))
  }
  fsr_by_lfsr
}


# LFSR for MASH -----------------------------------------------------
mash_lfsr &lt;- function(m, true_Y, step=0.05) {
  lfsr &lt;- get_lfsr(m)
  signs &lt;- get_pm(m) &gt; 0
  correct_signs &lt;- true_Y &gt; 0
  gotitright &lt;- signs == correct_signs

  nsteps &lt;- floor(.5 / step)
  fsr_by_lfsr &lt;- rep(0, nsteps)
  for (k in 1:nsteps) {
    idx &lt;- (lfsr &gt;= (step * (k - 1)) &amp; lfsr &lt; (step * k))
    fsr_by_lfsr[k] &lt;- ifelse(sum(idx) == 0, 0,
                             1 - sum(gotitright[idx]) / sum(idx))
  }
  fsr_by_lfsr
}</code></pre>
<p>Augmented FLASH simulation.</p>
<pre class="r"><code># Simulate from FLASH model -----------------------------------------
n &lt;- 1000
p &lt;- 10
flash_factors &lt;- 5

# Use one factor of all ones and one more interesting factor
nfactors &lt;- 2
k &lt;- p + nfactors
ff &lt;- matrix(0, nrow=k, ncol=p)
ff[1, ] &lt;- rep(10, p)
ff[2, ] &lt;- c(seq(10, 2, by=-2), rep(0, p - 5))
diag(ff[3:k, ]) &lt;- 3
ll &lt;- matrix(rnorm(n * k), nrow=n, ncol=k)
true_flash_Y &lt;- ll %*% ff
flash_Y &lt;- true_flash_Y + rnorm(n*p)
# RESULTS
flash_res &lt;- flash_v_mash(flash_Y, true_flash_Y, flash_factors)</code></pre>
<pre><code>fitting factor/loading 1</code></pre>
<pre><code>fitting factor/loading 2</code></pre>
<pre><code>fitting factor/loading 3</code></pre>
<pre><code>fitting factor/loading 4</code></pre>
<pre><code>fitting factor/loading 5</code></pre>
<pre><code> - Computing 1000 x 463 likelihood matrix.
 - Likelihood calculations took 0.12 seconds.
 - Fitting model with 463 mixture components.
 - Model fitting took 0.26 seconds.
 - Computing posterior matrices.
 - Computation allocated took 0.03 seconds.</code></pre>
<p>FLASH simulation.</p>
<pre class="r"><code># Simulate from basic FLASH model -----------------------------------
ff &lt;- ff[1:nfactors, ]
ll &lt;- matrix(rnorm(n * nfactors), nrow=n, ncol=nfactors)
true_basic_Y &lt;- ll %*% ff
basic_Y &lt;- true_basic_Y + rnorm(n*p)
# RESULTS
basic_res &lt;- flash_v_mash(basic_Y, true_basic_Y, flash_factors)</code></pre>
<pre><code>fitting factor/loading 1</code></pre>
<pre><code>fitting factor/loading 2</code></pre>
<pre><code>fitting factor/loading 3</code></pre>
<pre><code> - Computing 1000 x 463 likelihood matrix.
 - Likelihood calculations took 0.12 seconds.
 - Fitting model with 463 mixture components.</code></pre>
<pre><code>Warning in REBayes::KWDual(A, rep(1, k), normalize(w), control = control): estimated mixing distribution has some negative values:
               consider reducing rtol</code></pre>
<pre><code>Warning in mixIP(matrix_lik = structure(c(4.36775456036503e-15, 0, 0,
0, : Optimization step yields mixture weights that are either too small,
or negative; weights have been corrected and renormalized after the
optimization.</code></pre>
<pre><code> - Model fitting took 0.24 seconds.
 - Computing posterior matrices.
 - Computation allocated took 0.29 seconds.</code></pre>
<p>MASH simulation.</p>
<pre class="r"><code># Simulate from MASH model ------------------------------------------
Sigma &lt;- list()
Sigma[[1]] &lt;- matrix(1, nrow=p, ncol=p)
Sigma[[2]] &lt;- matrix(0, nrow=p, ncol=p)
for (i in 1:p) {
  for (j in 1:p) {
    Sigma[[2]][i, j] &lt;- max(1 - abs(i - j) / 4, 0)
  }
}
for (k in 1:p) {
  Sigma[[k + 2]] &lt;- matrix(0, nrow=p, ncol=p)
  Sigma[[k + 2]][k, k] &lt;- 1
}
which_sigma &lt;- sample(1:12, 1000, T, prob=c(.3, .3, rep(.4/p, p)))
true_mash_Y &lt;- matrix(0, nrow=n, ncol=p)
for (i in 1:n) {
  true_mash_Y[i, ] &lt;- 5*mvrnorm(1, rep(0, p), Sigma[[which_sigma[i]]])
}
mash_Y &lt;- true_mash_Y + rnorm(n * p)
# RESULTS
mash_res &lt;- flash_v_mash(mash_Y, true_mash_Y, flash_factors)</code></pre>
<pre><code>fitting factor/loading 1</code></pre>
<pre><code>fitting factor/loading 2</code></pre>
<pre><code>fitting factor/loading 3</code></pre>
<pre><code>fitting factor/loading 4</code></pre>
<pre><code>fitting factor/loading 5</code></pre>
<pre><code> - Computing 1000 x 400 likelihood matrix.
 - Likelihood calculations took 0.11 seconds.
 - Fitting model with 400 mixture components.
 - Model fitting took 0.58 seconds.
 - Computing posterior matrices.
 - Computation allocated took 0.03 seconds.</code></pre>
</div>
<div id="summary" class="section level2">
<h2>Summary</h2>
<p>In each case below, I follow the vignettes to produce a MASH fit (I use both canonical and data-driven covariance matrices). I fit a FLASH object (fixing the standard errors) by adding up to 10 factors greedily, then adding <span class="math inline">\(p\)</span> fixed one-hot vectors, and finally backfitting.</p>
<p>The two fits perform similarly. The MASH fit does somewhat better on data generated from the MASH model; more surprisingly, it performs comparably to FLASH on data generated from both the standard two-factor FLASH model. Both do poorly on the “augmented FLASH model” (described below), with MSEs near 1 (which would be obtained by simply using <span class="math inline">\(Y\)</span> as an estimate).</p>
</div>
<div id="flash-model" class="section level2">
<h2>Flash Model</h2>
<p>First I simulate from the basic FLASH model <span class="math inline">\(Y = LF + E\)</span> with <span class="math inline">\(E_{ij} \sim N(0, 1)\)</span>. Here, <span class="math inline">\(Y \in \mathbb{R}^{1000 \times 10}\)</span>, <span class="math inline">\(L \in \mathbb{R}^{1000 \times 2}\)</span> has i.i.d. <span class="math inline">\(N(0, 1)\)</span> entries, and <span class="math inline">\(F\)</span> is as follows:</p>
<pre><code>     [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9] [,10]
[1,]   10   10   10   10   10   10   10   10   10    10
[2,]   10    8    6    4    2    0    0    0    0     0</code></pre>
<p><strong>The MSE of the FLASH fit is 0.2, vs. 0.21 for the MASH fit. The proportion of 95% confidence intervals that contain the true value <span class="math inline">\(LF_{ij}\)</span> is 0.94 for FLASH and 0.96 for MASH.</strong> The true false sign rate vs lfsr appears as follows:</p>
<p><img src="figure/mashvflash.Rmd/lfsr1-1.png" width="672" style="display: block; margin: auto;" /></p>
<details> <summary><em>Expand here to see past versions of lfsr1-1.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/624de1dfb3db11bdc40a4a3d946a8fbd7abf40ce/docs/figure/mashvflash.Rmd/lfsr1-1.png" target="_blank">624de1d</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</td>
</tr>
<tr>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/9f81f0649020df46c440e91d61356176321aa930/docs/figure/mashvflash.Rmd/lfsr1-1.png" target="_blank">9f81f06</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</td>
</tr>
<tr>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/17f4dfb20b57e7e3b41a94a7b2dce787af325b3a/docs/figure/mashvflash.Rmd/lfsr1-1.png" target="_blank">17f4dfb</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</td>
</tr>
<tr>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/6508f2be3666427c3f8bd6c348e59ab7b9355565/docs/figure/mashvflash.Rmd/lfsr1-1.png" target="_blank">6508f2b</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-09
</td>
</tr>
<tr>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/861d4b3595125a0b8a304aa13375835acf4c8505/docs/figure/mashvflash.Rmd/lfsr1-1.png" target="_blank">861d4b3</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-09
</td>
</tr>
<tr>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/a76f3cf9f3dad726abb36bc409df06e9c31f57d2/docs/figure/mashvflash.Rmd/lfsr1-1.png" target="_blank">a76f3cf</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-09
</td>
</tr>
</tbody>
</table>
<p></details></p>
<p>The FLASH fit took 0.61 s. The MASH fit took 8.86 s.</p>
</div>
<div id="augmented-flash-model" class="section level2">
<h2>Augmented Flash Model</h2>
<p>Next I simulate from the “augmented” FLASH model <span class="math display">\[ Y = L \begin{pmatrix} F \\ 3I_{10} \end{pmatrix} + E \]</span> with <span class="math inline">\(F\)</span> as above.</p>
<p><strong>The MSE of the FLASH fit is 0.93, vs. 1.05 for the MASH fit. The proportion of 95% confidence intervals that contain the true value is 0.94 for FLASH and 0.93 for MASH.</strong> The true false sign rate vs lfsr appears as follows:</p>
<p><img src="figure/mashvflash.Rmd/lfsr2-1.png" width="672" style="display: block; margin: auto;" /></p>
<details> <summary><em>Expand here to see past versions of lfsr2-1.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/a1180b51f9cae5996e43a7bb368317bad76cf981/docs/figure/mashvflash.Rmd/lfsr2-1.png" target="_blank">a1180b5</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</td>
</tr>
<tr>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/624de1dfb3db11bdc40a4a3d946a8fbd7abf40ce/docs/figure/mashvflash.Rmd/lfsr2-1.png" target="_blank">624de1d</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</td>
</tr>
<tr>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/9f81f0649020df46c440e91d61356176321aa930/docs/figure/mashvflash.Rmd/lfsr2-1.png" target="_blank">9f81f06</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</td>
</tr>
<tr>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/17f4dfb20b57e7e3b41a94a7b2dce787af325b3a/docs/figure/mashvflash.Rmd/lfsr2-1.png" target="_blank">17f4dfb</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</td>
</tr>
<tr>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/6508f2be3666427c3f8bd6c348e59ab7b9355565/docs/figure/mashvflash.Rmd/lfsr2-1.png" target="_blank">6508f2b</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-09
</td>
</tr>
<tr>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/861d4b3595125a0b8a304aa13375835acf4c8505/docs/figure/mashvflash.Rmd/lfsr2-1.png" target="_blank">861d4b3</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-09
</td>
</tr>
<tr>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/a76f3cf9f3dad726abb36bc409df06e9c31f57d2/docs/figure/mashvflash.Rmd/lfsr2-1.png" target="_blank">a76f3cf</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-09
</td>
</tr>
</tbody>
</table>
<p></details></p>
<p>The FLASH fit took 18.36 s. The MASH fit took 3.57 s.</p>
</div>
<div id="mash-model" class="section level2">
<h2>MASH Model</h2>
<p>Finally I simulate from the MASH model <span class="math display">\[ X \sim \sum \pi_i N(0, \Sigma_i),\ Y = X + E \]</span> with <span class="math inline">\(E_{ij} \sim N(0, 1)\)</span>. I set <span class="math inline">\(\Sigma_1\)</span> to be the all ones matrix, <span class="math inline">\(\Sigma_2\)</span> to be a banded covariance matrix with non-zero entries on the first three off-diagonals, and <span class="math inline">\(\Sigma_3\)</span> through <span class="math inline">\(\Sigma_{12}\)</span> to have a single non-zero entry (corresponding to tissue-specific effects). <span class="math inline">\(\pi\)</span> is set to <span class="math inline">\((0.3, 0.3, 0.04, 0.04, \ldots, 0.04)\)</span>.</p>
<p><strong>The MSE of the FLASH fit is 0.56, vs. 0.43 for the MASH fit. The proportion of 95% confidence intervals that contain the true value is 0.9 for FLASH and 0.94 for MASH.</strong> The true false sign rate vs lfsr appears as follows:</p>
<p><img src="figure/mashvflash.Rmd/lfsr3-1.png" width="672" style="display: block; margin: auto;" /></p>
<details> <summary><em>Expand here to see past versions of lfsr3-1.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/a1180b51f9cae5996e43a7bb368317bad76cf981/docs/figure/mashvflash.Rmd/lfsr3-1.png" target="_blank">a1180b5</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</td>
</tr>
<tr>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/624de1dfb3db11bdc40a4a3d946a8fbd7abf40ce/docs/figure/mashvflash.Rmd/lfsr3-1.png" target="_blank">624de1d</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</td>
</tr>
<tr>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/9f81f0649020df46c440e91d61356176321aa930/docs/figure/mashvflash.Rmd/lfsr3-1.png" target="_blank">9f81f06</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-11
</td>
</tr>
<tr>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/6508f2be3666427c3f8bd6c348e59ab7b9355565/docs/figure/mashvflash.Rmd/lfsr3-1.png" target="_blank">6508f2b</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-09
</td>
</tr>
<tr>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/861d4b3595125a0b8a304aa13375835acf4c8505/docs/figure/mashvflash.Rmd/lfsr3-1.png" target="_blank">861d4b3</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-09
</td>
</tr>
<tr>
<td style="text-align:left;">
<a href="https://github.com/willwerscheid/MASHvFLASH/blob/a76f3cf9f3dad726abb36bc409df06e9c31f57d2/docs/figure/mashvflash.Rmd/lfsr3-1.png" target="_blank">a76f3cf</a>
</td>
<td style="text-align:left;">
Jason Willwerscheid
</td>
<td style="text-align:left;">
2018-06-09
</td>
</tr>
</tbody>
</table>
<p></details></p>
<p>The FLASH fit took 27.86 s. The MASH fit took 3.38 s.</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 Sierra 10.12.6

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] MASS_7.3-48  mashr_0.2-7  ashr_2.2-7   flashr_0.5-8

loaded via a namespace (and not attached):
 [1] Rcpp_0.12.16             pillar_1.2.1
 [3] plyr_1.8.4               compiler_3.4.3
 [5] git2r_0.21.0             workflowr_1.0.1
 [7] R.methodsS3_1.7.1        R.utils_2.6.0
 [9] iterators_1.0.9          tools_3.4.3
[11] testthat_2.0.0           digest_0.6.15
[13] tibble_1.4.2             evaluate_0.10.1
[15] memoise_1.1.0            gtable_0.2.0
[17] lattice_0.20-35          rlang_0.2.0
[19] Matrix_1.2-12            foreach_1.4.4
[21] commonmark_1.4           yaml_2.1.17
[23] parallel_3.4.3           mvtnorm_1.0-7
[25] ebnm_0.1-11              withr_2.1.1.9000
[27] stringr_1.3.0            roxygen2_6.0.1.9000
[29] xml2_1.2.0               knitr_1.20
[31] REBayes_1.2              devtools_1.13.4
[33] rprojroot_1.3-2          grid_3.4.3
[35] R6_2.2.2                 rmarkdown_1.8
[37] rmeta_3.0                ggplot2_2.2.1
[39] magrittr_1.5             whisker_0.3-2
[41] backports_1.1.2          scales_0.5.0
[43] codetools_0.2-15         htmltools_0.3.6
[45] assertthat_0.2.0         softImpute_1.4
[47] colorspace_1.3-2         stringi_1.1.6
[49] Rmosek_7.1.3             lazyeval_0.2.1
[51] munsell_0.4.3            doParallel_1.0.11
[53] pscl_1.5.2               truncnorm_1.0-8
[55] SQUAREM_2017.10-1        ExtremeDeconvolution_1.3
[57] 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>