Skip to content

Commit

Permalink
README
Browse files Browse the repository at this point in the history
  • Loading branch information
@Mahsa-Ehsanifard
Mahsa-Ehsanifard committed Aug 13, 2024
1 parent e82b07c commit a55dd25
Show file tree
Hide file tree
Showing 2 changed files with 132 additions and 0 deletions.
63 changes: 63 additions & 0 deletions README.html
View file
Original file line number Diff line number Diff line change
Expand Up @@ -421,6 +421,69 @@ <h3>Detecting the outlier samples</h3>
<li>The outlier samples are removed from the main matrix.</li>
</ul>
</div>
<div id="loading-traits" class="section level3">
<h3>loading traits</h3>
<p>We need a table of <em>trait</em> information such as <em>clinical
data, molecular characteristics, or phenotypic and genotypic
features</em> to analyze the correlation and relationship between traits
and gene expressions. In this step, it is needed to investigate the
relationships of traits with gene modules to identify hub genes
correlated strongly with an important features of samples.</p>
<ul>
<li><p>Trait could be the features or characteristics of genes such as
regulation levels.</p></li>
<li><p>Clinical data includes staging, mutations, molecular or cellular
features, etc…</p></li>
</ul>
</div>
<div id="choose-a-set-of-soft-thresholding-power" class="section level3">
<h3>Choose a set of soft thresholding power</h3>
<p>Choosing a <strong>soft power (β)</strong> is an important step to
detect modules.</p>
<ul>
<li><p>power number is a critical index to identify gene module
packing</p></li>
<li><p>β parameter will be to calculate our adjacency matrix.</p></li>
<li><p>The <code>pickSoftThreshold</code> function calculates multiple
networks all based on different β values and returns a data frame with
the <strong>R2</strong> values for the networks <strong>scale-free
topology</strong> model fit as well as the <strong>mean
connectivity</strong> measures.</p></li>
</ul>
<p><code>{r} pickSoftThreshold(matrix)</code></p>
<pre><code> Power SFT.R.sq slope truncated.R.sq mean.k. median.k. max.k.
## 1 1 0.0278 0.345 0.456 747.00 762.0000 1210.0
## 2 2 0.1260 -0.597 0.843 254.00 251.0000 574.0
## 3 3 0.3400 -1.030 0.972 111.00 102.0000 324.0
## 4 4 0.5060 -1.420 0.973 56.50 47.2000 202.0
## 5 5 0.6810 -1.720 0.940 32.20 25.1000 134.0
## 6 6 0.9020 -1.500 0.962 19.90 14.5000 94.8
## 7 7 0.9210 -1.670 0.917 13.20 8.6800 84.1
## 8 8 0.9040 -1.720 0.876 9.25 5.3900 76.3
## 9 9 0.8590 -1.700 0.836 6.80 3.5600 70.5
## 10 10 0.8330 -1.660 0.831 5.19 2.3800 65.8
## 11 12 0.8530 -1.480 0.911 3.33 1.1500 58.1
## 12 14 0.8760 -1.380 0.949 2.35 0.5740 51.9
## 13 16 0.9070 -1.300 0.970 1.77 0.3090 46.8
## 14 18 0.9120 -1.240 0.973 1.39 0.1670 42.5
## 15 20 0.9310 -1.210 0.977 1.14 0.0951 38.7</code></pre>
<p>Plot the R2 values as a function of the soft thresholds</p>
<blockquote>
<p>We should be <em>maximizing</em> the R2 (β) value and
<em>minimizing</em> mean connectivity.</p>
</blockquote>
<p><code>{r} par(mfrow=c(1,2)) plot(sft$fitIndices[,1], -sign(sft$fitIndices[,3])*sft$fitIndices[,2], xlab=&quot;Soft Threshold (power)&quot;,ylab=&quot;Scale Free Topology Model Fit, signed Rˆ2&quot;,type=&quot;n&quot;,main=paste(&quot;Scale independence&quot;)) text(sft$fitIndices[,1], -sign(sft$fitIndices[,3])*sft$fitIndices[,2], labels=powers,col=&quot;red&quot;) abline(h=0.80,col=&quot;red&quot;) plot(sft$fitIndices[,1],sft$fitIndices[,5],type=&quot;n&quot;, xlab=&quot;Soft Threshold (power)&quot;,ylab=&quot;Mean Connectivity&quot;, main=paste(&quot;Mean connectivity&quot;)) text(sft$fitIndices[,1],sft$fitIndices[,5],labels=powers, col=&quot;red&quot;)</code></p>
<ul>
<li>We can determine the soft power threshold which is a number as it is
the β that retains the <em>highest</em> mean connectivity (above zero)
while reaching an R2 value above <strong>0.80</strong>.</li>
</ul>
<blockquote>
<p>NOTE: the higher the value, the stronger the connection strength will
be of highly correlated gene expression profiles and the more devalued
low correlations will be.</p>
</blockquote>
</div>
</div>
</div>

Expand Down
69 changes: 69 additions & 0 deletions README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -75,6 +75,75 @@ keepsample <- clust==1

* The outlier samples are removed from the main matrix.

### loading traits

We need a table of *trait* information such as *clinical data, molecular characteristics, or phenotypic and genotypic features* to analyze the correlation and relationship between traits and gene expressions. In this step, it is needed to investigate the relationships of traits with gene modules to identify hub genes correlated strongly with an important features of samples.

* Trait could be the features or characteristics of genes such as regulation levels.

* Clinical data includes staging, mutations, molecular or cellular features, etc...

### Choose a set of soft thresholding power

Choosing a **soft power (β)** is an important step to detect modules.

* power number is a critical index to identify gene module packing

* β parameter will be to calculate our adjacency matrix.

* The `pickSoftThreshold` function calculates multiple networks all based on different β values and returns a data frame with the **R2** values for the networks **scale-free topology** model fit as well as the **mean connectivity** measures.

```{r}
pickSoftThreshold(matrix)
```

```
Power SFT.R.sq slope truncated.R.sq mean.k. median.k. max.k.
## 1 1 0.0278 0.345 0.456 747.00 762.0000 1210.0
## 2 2 0.1260 -0.597 0.843 254.00 251.0000 574.0
## 3 3 0.3400 -1.030 0.972 111.00 102.0000 324.0
## 4 4 0.5060 -1.420 0.973 56.50 47.2000 202.0
## 5 5 0.6810 -1.720 0.940 32.20 25.1000 134.0
## 6 6 0.9020 -1.500 0.962 19.90 14.5000 94.8
## 7 7 0.9210 -1.670 0.917 13.20 8.6800 84.1
## 8 8 0.9040 -1.720 0.876 9.25 5.3900 76.3
## 9 9 0.8590 -1.700 0.836 6.80 3.5600 70.5
## 10 10 0.8330 -1.660 0.831 5.19 2.3800 65.8
## 11 12 0.8530 -1.480 0.911 3.33 1.1500 58.1
## 12 14 0.8760 -1.380 0.949 2.35 0.5740 51.9
## 13 16 0.9070 -1.300 0.970 1.77 0.3090 46.8
## 14 18 0.9120 -1.240 0.973 1.39 0.1670 42.5
## 15 20 0.9310 -1.210 0.977 1.14 0.0951 38.7
```

Plot the R2 values as a function of the soft thresholds

> We should be *maximizing* the R2 (β) value and *minimizing* mean connectivity.
```{r}
par(mfrow=c(1,2))
plot(sft$fitIndices[,1],
-sign(sft$fitIndices[,3])*sft$fitIndices[,2],
xlab="Soft Threshold (power)",ylab="Scale Free Topology
Model Fit,
signed Rˆ2",type="n",main=paste("Scale independence"))
text(sft$fitIndices[,1],
-sign(sft$fitIndices[,3])*sft$fitIndices[,2],
labels=powers,col="red")
abline(h=0.80,col="red")
plot(sft$fitIndices[,1],sft$fitIndices[,5],type="n",
xlab="Soft Threshold (power)",ylab="Mean Connectivity",
main=paste("Mean connectivity"))
text(sft$fitIndices[,1],sft$fitIndices[,5],labels=powers,
col="red")
```

* We can determine the soft power threshold which is a number as it is the β that retains the *highest* mean connectivity (above zero) while reaching an R2 value above **0.80**.

> NOTE: the higher the value, the stronger the connection strength will be of highly correlated gene expression profiles and the more devalued low correlations will be.





Expand Down

0 comments on commit a55dd25

Please sign in to comment.