<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta http-equiv="X-UA-Compatible" content="IE=Edge" />
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<title>The Maui Class — 0.1 documentation</title>
<link rel="stylesheet" href="_static/alabaster.css" type="text/css" />
<link rel="stylesheet" href="_static/pygments.css" type="text/css" />
<script type="text/javascript" id="documentation_options" data-url_root="./" src="_static/documentation_options.js"></script>
<script type="text/javascript" src="_static/jquery.js"></script>
<script type="text/javascript" src="_static/underscore.js"></script>
<script type="text/javascript" src="_static/doctools.js"></script>
<script async="async" type="text/javascript" src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.1/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
<link rel="shortcut icon" href="_static/icons8-beach-30.png"/>
<link rel="index" title="Index" href="genindex.html" />
<link rel="search" title="Search" href="search.html" />
<link rel="next" title="Maui Utilities" href="utils.html" />
<link rel="prev" title="Filtering and Merging latent factors" href="filtering-and-merging-latent-factors.html" />
<link rel="stylesheet" href="_static/custom.css" type="text/css" />
<meta name="viewport" content="width=device-width, initial-scale=0.9, maximum-scale=0.9" />
</head><body>
<div class="document">
<div class="documentwrapper">
<div class="bodywrapper">
<div class="body" role="main">
<div class="section" id="the-maui-class">
<h1>The Maui Class<a class="headerlink" href="#the-maui-class" title="Permalink to this headline">¶</a></h1>
<dl class="class">
<dt id="maui.Maui">
<em class="property">class </em><code class="descclassname">maui.</code><code class="descname">Maui</code><span class="sig-paren">(</span><em>n_hidden=[1500], n_latent=80, batch_size=100, epochs=400, architecture='stacked', initial_beta_val=0, kappa=1.0, max_beta_val=1, learning_rate=0.0005, epsilon_std=1.0, batch_normalize_inputs=True, batch_normalize_intermediaries=True, batch_normalize_embedding=True, relu_intermediaries=True, relu_embedding=True, input_dim=None</em><span class="sig-paren">)</span><a class="reference internal" href="_modules/maui/model.html#Maui"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#maui.Maui" title="Permalink to this definition">¶</a></dt>
<dd><p>Maui (Multi-omics Autoencoder Integration) model.</p>
<p>Trains a variational autoencoder to find latent factors in multi-modal data.</p>
<dl class="docutils">
<dt>n_hidden: array (default [1500])</dt>
<dd>The sizes of the hidden layers of the autoencoder architecture.
Each element of the array specifies the number of nodes in successive
layers of the autoencoder</dd>
<dt>n_latent: int (default 80)</dt>
<dd>The size of the latent layer (number of latent features)</dd>
<dt>batch_size: int (default 100)</dt>
<dd>The size of the mini-batches used for training the network</dd>
<dt>epochs: int (default 400)</dt>
<dd>The number of epoches to use for training the network</dd>
<dt>architecture:</dt>
<dd>One of ‘stacked’ or ‘deep’. If ‘stacked’, will use a stacked VAE model, where
the intermediate layers are also variational. If ‘deep’, will train a deep VAE
where the intermediate layers are regular (ReLU) units, and only the middle
(latent) layer is variational.</dd>
</dl>
<dl class="method">
<dt id="maui.Maui.c_index">
<code class="descname">c_index</code><span class="sig-paren">(</span><em>survival, clinical_only=True, duration_column='duration', observed_column='observed', cox_penalties=[0.1, 1, 10, 100, 1000, 10000], cv_folds=5, sel_clin_alpha=0.05, sel_clin_penalty=0</em><span class="sig-paren">)</span><a class="reference internal" href="_modules/maui/model.html#Maui.c_index"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#maui.Maui.c_index" title="Permalink to this definition">¶</a></dt>
<dd><p>Compute’s Harrell’s c-Index for a Cox Proportional Hazards regression modeling
survival by the latent factors in z.</p>
<p>z: pd.DataFrame (n_samples, n_latent factors)
survival: pd.DataFrame of survival information and relevant covariates</p>
<blockquote>
<div>(such as sex, age at diagnosis, or tumor stage)</div></blockquote>
<dl class="docutils">
<dt>clinical_only: Compute the c-Index for a model containing only</dt>
<dd>individually clinically relevant latent factors
(see <code class="docutils literal notranslate"><span class="pre">select_clinical_factors</span></code>)</dd>
<dt>duration_column: the name of the column in <code class="docutils literal notranslate"><span class="pre">survival</span></code> containing the</dt>
<dd>duration (time between diagnosis and death or last followup)</dd>
<dt>observed_column: the name of the column in <code class="docutils literal notranslate"><span class="pre">survival</span></code> containing</dt>
<dd>indicating whether time of death is known</dd>
<dt>cox_penalties: penalty coefficient in Cox PH solver (see <code class="docutils literal notranslate"><span class="pre">lifelines.CoxPHFitter</span></code>)</dt>
<dd>to try. Returns the best c given by the different penalties
(by cross-validation)</dd>
</dl>
<p>cv_folds: number of cross-validation folds to compute C
sel_clin_penalty: CPH penalizer to use when selecting clinical factors
sel_clin_alpha: significance level when selecting clinical factors</p>
<dl class="docutils">
<dt>cs: array, Harrell’s c-Index, an auc-like metric for survival prediction accuracy.</dt>
<dd>one value per cv_fold</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="maui.Maui.cluster">
<code class="descname">cluster</code><span class="sig-paren">(</span><em>k=None</em>, <em>optimal_k_method='ami'</em>, <em>optimal_k_range=range(3</em>, <em>10)</em>, <em>ami_y=None</em>, <em>kmeans_kwargs={'n_init': 1000</em>, <em>'n_jobs': 2}</em><span class="sig-paren">)</span><a class="reference internal" href="_modules/maui/model.html#Maui.cluster"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#maui.Maui.cluster" title="Permalink to this definition">¶</a></dt>
<dd><p>Cluster the samples using k-means based on the latent factors.</p>
<dl class="docutils">
<dt>k: optional, the number of clusters to find.</dt>
<dd>if not given, will attempt to find optimal k.</dd>
<dt>optimal_k_method: supported methods are ‘ami’ and ‘silhouette’. Otherwise, callable.</dt>
<dd>if ‘ami’, will pick K which gives the best AMI
(adjusted mutual information) with external labels.
if ‘silhouette’ will pick the K which gives the best
mean silhouette coefficient.
if callable, should have signature <code class="docutils literal notranslate"><span class="pre">scorer(yhat)</span></code>
and return a scalar score.</dd>
</dl>
<p>optimal_k_range: array-like, range of Ks to try to find optimal K among
ami_y: array-like (n_samples), the ground-truth labels to use</p>
<blockquote>
<div>when picking K by “best AMI against ground-truth” method.</div></blockquote>
<p>kmeans_kwargs: optional, kwargs for initialization of sklearn.cluster.KMeans</p>
<p>yhat: Series (n_samples) cluster labels for each sample</p>
</dd></dl>
<dl class="method">
<dt id="maui.Maui.compute_auc">
<code class="descname">compute_auc</code><span class="sig-paren">(</span><em>y</em>, <em>**kwargs</em><span class="sig-paren">)</span><a class="reference internal" href="_modules/maui/model.html#Maui.compute_auc"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#maui.Maui.compute_auc" title="Permalink to this definition">¶</a></dt>
<dd><p>Compute area under the ROC curve for predicting the labels in y using the
latent features previously inferred.</p>
<p>y: labels to predict
<a href="#id1"><span class="problematic" id="id2">**</span></a>kwargs: arguments for <code class="docutils literal notranslate"><span class="pre">compute_roc</span></code></p>
<p>aucs: pd.Series, auc per class as well as mean</p>
</dd></dl>
<dl class="method">
<dt id="maui.Maui.compute_roc">
<code class="descname">compute_roc</code><span class="sig-paren">(</span><em>y</em>, <em>**kwargs</em><span class="sig-paren">)</span><a class="reference internal" href="_modules/maui/model.html#Maui.compute_roc"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#maui.Maui.compute_roc" title="Permalink to this definition">¶</a></dt>
<dd><p>Compute Receiver Operating Characteristics curve for SVM prediction
of labels <code class="docutils literal notranslate"><span class="pre">y</span></code> from the latent factors. Computes both the ROC curves
(true positive rate, true negative rate), and the area under the roc (auc).
ROC and auROC computed for each class (the classes are inferred from <code class="docutils literal notranslate"><span class="pre">y</span></code>),
as well as a “mean” ROC, computed by averaging the class ROCs. Only samples
in the index of <code class="docutils literal notranslate"><span class="pre">y</span></code> will be considered.</p>
<p>y: array-like (n_samples,), the labels of the samples to predict
<a href="#id3"><span class="problematic" id="id4">**</span></a>kwargs: arguments for <code class="docutils literal notranslate"><span class="pre">utils.compute_roc</span></code></p>
<dl class="docutils">
<dt>roc_curves: dict, one key per class as well as “mean”, each value is a dataframe</dt>
<dd>containing the tpr (true positive rate) and fpr (falce positive rate)
defining that class (or the mean) ROC.</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="maui.Maui.drop_unexplanatory_factors">
<code class="descname">drop_unexplanatory_factors</code><span class="sig-paren">(</span><em>threshold=0.02</em><span class="sig-paren">)</span><a class="reference internal" href="_modules/maui/model.html#Maui.drop_unexplanatory_factors"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#maui.Maui.drop_unexplanatory_factors" title="Permalink to this definition">¶</a></dt>
<dd><p>Drops factors which have a low R^2 score in a univariate linear model
predicting the features <cite>x</cite> from a column of the latent factors <cite>z</cite>.</p>
<dl class="docutils">
<dt>threshold: threshold for R^2, latent factors below this threshold</dt>
<dd>are dropped.</dd>
</dl>
<dl class="docutils">
<dt>z_filt: (n_samples, n_factors) DataFrame of latent factor values,</dt>
<dd>with only those columns from the input <cite>z</cite> which have an R^2
above the threshold when using that column as an input
to a linear model predicting <cite>x</cite>.</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="maui.Maui.fit">
<code class="descname">fit</code><span class="sig-paren">(</span><em>X</em>, <em>y=None</em>, <em>X_validation=None</em><span class="sig-paren">)</span><a class="reference internal" href="_modules/maui/model.html#Maui.fit"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#maui.Maui.fit" title="Permalink to this definition">¶</a></dt>
<dd><p>Train autoencoder model</p>
<dl class="docutils">
<dt>X: dict with multi-modal dataframes, containing training data, e.g.</dt>
<dd>{‘mRNA’: df1, ‘SNP’: df2},
df1, df2, etc. are (n_features, n_samples) pandas.DataFrame’s.
The sample names must match, the feature names need not.</dd>
<dt>X_validation: optional, dict with multi-modal dataframes, containing validation data</dt>
<dd>will be used to compute validation loss under training</dd>
</dl>
<p>y: Not used.</p>
<p>self : Maui object</p>
</dd></dl>
<dl class="method">
<dt id="maui.Maui.fit_transform">
<code class="descname">fit_transform</code><span class="sig-paren">(</span><em>X</em>, <em>y=None</em>, <em>X_validation=None</em>, <em>encoder='mean'</em><span class="sig-paren">)</span><a class="reference internal" href="_modules/maui/model.html#Maui.fit_transform"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#maui.Maui.fit_transform" title="Permalink to this definition">¶</a></dt>
<dd><p>Train autoencoder model, and return the latent factor representation
of the data X.</p>
<dl class="docutils">
<dt>X: dict with multi-modal dataframes, containing training data, e.g.</dt>
<dd>{‘mRNA’: df1, ‘SNP’: df2},
df1, df2, etc. are (n_samples, n_features) pandas.DataFrame’s.
The sample names must match, the feature names need not.</dd>
<dt>X_validation: optional, dict with multi-modal dataframes, containing validation data</dt>
<dd>will be used to compute validation loss under training</dd>
</dl>
<p>y: Not used.</p>
<dl class="docutils">
<dt>z: DataFrame (n_samples, n_latent_factors)</dt>
<dd>Latent factors representation of the data X.</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="maui.Maui.get_linear_weights">
<code class="descname">get_linear_weights</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="_modules/maui/model.html#Maui.get_linear_weights"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#maui.Maui.get_linear_weights" title="Permalink to this definition">¶</a></dt>
<dd><p>Get linear model coefficients obtained from fitting linear models
predicting feature values from latent factors. One model is fit per latent
factor, and the coefficients are stored in the matrix.</p>
<dl class="docutils">
<dt>W: (n_features, n_latent_factors) DataFrame</dt>
<dd>w_{ij} is the coefficient associated with feature <cite>i</cite> in a linear model
predicting it from latent factor <cite>j</cite>.</dd>
</dl>
</dd></dl>
<dl class="staticmethod">
<dt id="maui.Maui.load">
<em class="property">static </em><code class="descname">load</code><span class="sig-paren">(</span><em>dir</em><span class="sig-paren">)</span><a class="reference internal" href="_modules/maui/model.html#Maui.load"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#maui.Maui.load" title="Permalink to this definition">¶</a></dt>
<dd><p>Load a maui model from disk, which was previously saved using <code class="docutils literal notranslate"><span class="pre">save()</span></code></p>
<p>dir: The directory from which to load the maui model</p>
<p>maui_model: a maui model that was previously saved to disk</p>
</dd></dl>
<dl class="method">
<dt id="maui.Maui.merge_similar_latent_factors">
<code class="descname">merge_similar_latent_factors</code><span class="sig-paren">(</span><em>distance_in='z'</em>, <em>distance_metric='correlation'</em>, <em>linkage_method='complete'</em>, <em>distance_threshold=0.17</em>, <em>merge_fn=<function mean></em>, <em>plot_dendrogram=True</em>, <em>plot_dendro_ax=None</em><span class="sig-paren">)</span><a class="reference internal" href="_modules/maui/model.html#Maui.merge_similar_latent_factors"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#maui.Maui.merge_similar_latent_factors" title="Permalink to this definition">¶</a></dt>
<dd><p>Merge latent factorz in z whose distance is below a certain threshold.
Used to squeeze down latent factor representations if there are many co-linear
latent factors.</p>
<dl class="docutils">
<dt>distance_in: If ‘z’, latent factors will be merged based on their distance</dt>
<dd>to each other in ‘z’. If ‘w’, favtors will be merged based
on their distance in ‘w’ (see <a class="reference internal" href="#maui.Maui.get_linear_weights" title="maui.Maui.get_linear_weights"><code class="xref py py-func docutils literal notranslate"><span class="pre">get_linear_weights()</span></code></a>)</dd>
<dt>distance_metric: The distance metric based on which to merge latent factors.</dt>
<dd>One which is supported by <code class="xref py py-func docutils literal notranslate"><span class="pre">scipy.spatial.distance.pdist()</span></code></dd>
<dt>linkage_method: The linkage method used to cluster latent factors. One which</dt>
<dd>is supported by <code class="xref py py-func docutils literal notranslate"><span class="pre">scipy.cluster.hierarchy.linkage()</span></code>.</dd>
<dt>distance_threshold: Latent factors with distance below this threshold</dt>
<dd>will be merged</dd>
<dt>merge_fn: Function used to determine value of merged latent factor.</dt>
<dd>The default is <code class="xref py py-func docutils literal notranslate"><span class="pre">numpy.mean()</span></code>, meaning the merged
latent factor will have the mean value of the inputs.</dd>
<dt>plot_dendrogram: Boolean. If true, a dendrogram will be plotted showing</dt>
<dd>which latent factors are merged and the threshold.</dd>
</dl>
<p>plot_dendro_ax: A matplotlib axis object to plot the dendrogram on (optional)</p>
<dl class="docutils">
<dt>z: (n_samples, n_factors) pd.DataFrame of latent factors</dt>
<dd>where some have been merged</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="maui.Maui.save">
<code class="descname">save</code><span class="sig-paren">(</span><em>destdir</em><span class="sig-paren">)</span><a class="reference internal" href="_modules/maui/model.html#Maui.save"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#maui.Maui.save" title="Permalink to this definition">¶</a></dt>
<dd><p>Save a maui model to disk, so that it may be reloaded later using <code class="docutils literal notranslate"><span class="pre">load()</span></code></p>
<p>destdir: destination directory in which to save model files</p>
</dd></dl>
<dl class="method">
<dt id="maui.Maui.select_clinical_factors">
<code class="descname">select_clinical_factors</code><span class="sig-paren">(</span><em>survival</em>, <em>duration_column='duration'</em>, <em>observed_column='observed'</em>, <em>alpha=0.05</em>, <em>cox_penalizer=0</em><span class="sig-paren">)</span><a class="reference internal" href="_modules/maui/model.html#Maui.select_clinical_factors"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#maui.Maui.select_clinical_factors" title="Permalink to this definition">¶</a></dt>
<dd><p>Select latent factors which are predictive of survival. This is
accomplished by fitting a Cox Proportional Hazards (CPH) model to each
latent factor, while controlling for known covariates, and only keeping
those latent factors whose coefficient in the CPH is nonzero (adjusted
p-value < alpha).</p>
<dl class="docutils">
<dt>survival: pd.DataFrame of survival information and relevant covariates</dt>
<dd>(such as sex, age at diagnosis, or tumor stage)</dd>
<dt>duration_column: the name of the column in <code class="docutils literal notranslate"><span class="pre">survival</span></code> containing the</dt>
<dd>duration (time between diagnosis and death or last followup)</dd>
<dt>observed_column: the name of the column in <code class="docutils literal notranslate"><span class="pre">survival</span></code> containing</dt>
<dd>indicating whether time of death is known</dd>
<dt>alpha: threshold for p-value of CPH coefficients to call a latent</dt>
<dd>factor clinically relevant (p < alpha)</dd>
</dl>
<p>cox_penalizer: penalty coefficient in Cox PH solver (see <code class="docutils literal notranslate"><span class="pre">lifelines.CoxPHFitter</span></code>)</p>
<dl class="docutils">
<dt>z_clinical: pd.DataFrame, subset of the latent factors which have been</dt>
<dd>determined to have clinical value (are individually predictive
of survival, controlling for covariates)</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="maui.Maui.transform">
<code class="descname">transform</code><span class="sig-paren">(</span><em>X</em>, <em>encoder='mean'</em><span class="sig-paren">)</span><a class="reference internal" href="_modules/maui/model.html#Maui.transform"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#maui.Maui.transform" title="Permalink to this definition">¶</a></dt>
<dd><p>Transform X into the latent space that was previously learned using
<cite>fit</cite> or <cite>fit_transform</cite>, and return the latent factor representation.</p>
<dl class="docutils">
<dt>X: dict with multi-modal dataframes, containing training data, e.g.</dt>
<dd>{‘mRNA’: df1, ‘SNP’: df2},
df1, df2, etc. are (n_features, n_samples) pandas.DataFrame’s.</dd>
<dt>encoder: the mode of the encoder to be used. one of ‘mean’ or ‘sample’,</dt>
<dd>where ‘mean’ indicates the encoder network only uses the mean
estimates for each successive layer. ‘sample’ indicates the
encoder should sample from the distribution specified from each
successive layer, and results in non-reproducible embeddings.</dd>
</dl>
<dl class="docutils">
<dt>z: DataFrame (n_samples, n_latent_factors)</dt>
<dd>Latent factors representation of the data X.</dd>
</dl>
</dd></dl>
</dd></dl>
</div>
</div>
</div>
</div>
<div class="sphinxsidebar" role="navigation" aria-label="main navigation">
<div class="sphinxsidebarwrapper">
<p class="logo"><a href="index.html">
<img class="logo" src="_static/hex-maui.png" alt="Logo"/>
</a></p>
<h1 class="logo"><a href="index.html"></a></h1>
<h3>Navigation</h3>
<ul class="current">
<li class="toctree-l1"><a class="reference internal" href="autoencoder-integration.html">Multi-modal Autoencoders</a></li>
<li class="toctree-l1"><a class="reference internal" href="data-normalization.html">Data and Normalization</a></li>
<li class="toctree-l1"><a class="reference internal" href="filtering-and-merging-latent-factors.html">Filtering and Merging latent factors</a></li>
<li class="toctree-l1 current"><a class="current reference internal" href="#">The Maui Class</a></li>
<li class="toctree-l1"><a class="reference internal" href="utils.html">Maui Utilities</a></li>
</ul>
<div class="relations">
<h3>Related Topics</h3>
<ul>
<li><a href="index.html">Documentation overview</a><ul>
<li>Previous: <a href="filtering-and-merging-latent-factors.html" title="previous chapter">Filtering and Merging latent factors</a></li>
<li>Next: <a href="utils.html" title="next chapter">Maui Utilities</a></li>
</ul></li>
</ul>
</div>
<div id="searchbox" style="display: none" role="search">
<h3>Quick search</h3>
<div class="searchformwrapper">
<form class="search" action="search.html" method="get">
<input type="text" name="q" />
<input type="submit" value="Go" />
<input type="hidden" name="check_keywords" value="yes" />
<input type="hidden" name="area" value="default" />
</form>
</div>
</div>
<script type="text/javascript">$('#searchbox').show(0);</script>
</div>
</div>
<div class="clearer"></div>
</div>
<div class="footer">
©2018, Jonathan Ronen, Altuna Akalin.
|
Powered by <a href="http://sphinx-doc.org/">Sphinx 1.8.2</a>
& <a href="https://github.com/bitprophet/alabaster">Alabaster 0.7.12</a>
|
<a href="_sources/maui.rst.txt"
rel="nofollow">Page source</a>
</div>
</body>
</html>
Datasets
Models
