Commit 0bfdbba9 authored by yannponty's avatar yannponty
Browse files

small stuff

parent 50ca25de
......@@ -468,7 +468,7 @@ It is important to recall that the set of positives and negatives is influenced
% The structural disruption is evaluated with Mutate and Map in the first column and \remu in the second.}
% \label{fig:auc}
%\end{figure}
\subsection{Evolutionary stabilization of {\itshape in vitro} disruptive mutations reveal molecular interfaces}
\subsection{Evolutionary stabilization of {\itshape in vitro} disruptive mutations reveal binding sites}
The ROC curves for the 4 PDB structures are shown on the same graph, for a specific \shape distance threshold $\delta$ and $\gamma$.
We illustrate in Fig.~\ref{fig:roc}a the results for the 5S RNA given $\delta$ at the $96^{\text{th}}$ percentile and a $\gamma$ of $23$.
......@@ -562,7 +562,7 @@ The poorest results are achieved in family RF01118. Interestingly, one of the co
These observations validate our methodology and at the same time justify the use of \shape data.
To complete this analysis, we also investigate the ratio and size of the overlap between the structurally-disruptive positions predicted using \remu and \shape experiments, for different percentiles. As indicated in Fig.~S4, we notice a clear linear decrease in the size of the overlap. At the $50^{\text{th}}$ percentile, the size of the intersection is cut by half. When combining the results of \remu and \shape experiments together, we quickly reach results that are almost as good as those obtained with \shape data alone, but then we also lose all specificity since the intersection sets are too small and appear to mainly identify mutations not found in the multiple sequence alignment. Those results are show in Fig.~S5.
To complete this analysis, we also investigate the ratio and size of the overlap between the structurally-disruptive positions predicted using \remu and \shape experiments, for different percentiles. As indicated in Fig.~S4, we notice a clear linear decrease in the size of the overlap. At the $50^{\text{th}}$ percentile, the size of the intersection is cut by half. When combining the results of \remu and \shape experiments together, we quickly reach results that are almost as good as those obtained with \shape data alone, but then we also lose all specificity since the intersection sets are too small and appear to mainly identify mutations not found in the multiple sequence alignment. Those results are shown in Fig.~S5.
These observations implies that although a theoretical model do capture part of the complexity of the structural conformation ensemble, it is currently too noisy to identify fine grain differences captured by the \shape experiments.
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment