cg
changeset 61:cb5eed6525f2
.
author | bshanks@bshanks.dyndns.org |
---|---|
date | Sun Apr 19 14:44:41 2009 -0700 (16 years ago) |
parents | 9381e0c1827f |
children | ecf330fcfba3 |
files | grant.doc grant.html grant.odt grant.pdf grant.txt |
line diff
1.1 Binary file grant.doc has changed
2.1 --- a/grant.html Sun Apr 19 14:29:24 2009 -0700
2.2 +++ b/grant.html Sun Apr 19 14:44:41 2009 -0700
2.3 @@ -103,12 +103,12 @@
2.4 pixels in the intersection of the two images, divided by the number of pixels in their union. Neither GeneAtlas nor EMAGE
2.5 allow one to search for combinations of genes that define a region in concert but not separately.
2.6 [10 ] describes AGEA, ”Anatomic Gene Expression Atlas”. AGEA has three components:
2.7 -* Gene Finder: The user selects a seed voxel and the system (1) chooses a cluster which includes the seed voxel, (2)
2.8 -yields a list of genes which are overexpressed in that cluster. (note: the ABA website also contains pre-prepared lists of
2.9 -overexpressed genes for selected structures)
2.10 -* Correlation: The user selects a seed voxel and the shows the user how much correlation there is between the gene
2.11 +∙Gene Finder: The user selects a seed voxel and the system (1) chooses a cluster which includes the seed voxel, (2)
2.12 +yields a list of genes which are overexpressed in that cluster. (note: the ABA website also contains pre-prepared lists
2.13 +of overexpressed genes for selected structures)
2.14 +∙Correlation: The user selects a seed voxel and the shows the user how much correlation there is between the gene
2.15 expression profile of the seed voxel and every other voxel.
2.16 -* Clusters: will be described later
2.17 +∙Clusters: will be described later
2.18 Gene Finder is different from our Aim 1 in at least three ways. First, Gene Finder finds only single genes, whereas we
2.19 will also look for combinations of genes. Second, gene finder can only use overexpression as a marker, whereas we will also
2.20 search for underexpression. Third, Gene Finder uses a simple pointwise score4, whereas we will also use geometric scores
2.21 @@ -351,7 +351,7 @@
2.22 One class of feature selection scoring method are those which calculate some sort of “match” between each gene image
2.23 and the target image. Those genes which match the best are good candidates for features.
2.24 One of the simplest methods in this class is to use correlation as the match score. We calculated the correlation between
2.25 -each gene and each cortical area. The top row of Figure shows the three genes most correlated with area SS.
2.26 +each gene and each cortical area. The top row of Figure 1 shows the three genes most correlated with area SS.
2.27 Conditional entropy An information-theoretic scoring method is to find features such that, if the features (gene
2.28 expression levels) are known, uncertainty about the target (the regional identity) is reduced. Entropy measures uncertainty,
2.29 so what we want is to find features such that the conditional distribution of the target has minimal entropy. The distribution
2.30 @@ -395,7 +395,7 @@
2.31 similar direction (because the borders are similar).
2.32 Gradient similarity provides information complementary to correlation
2.33 To show that gradient similarity can provide useful information that cannot be detected via pointwise analyses, consider
2.34 -Fig. . The top row of Fig. displays the 3 genes which most match area AUD, according to a pointwise method17. The
2.35 +Fig. 2. The top row of Fig. 2 displays the 3 genes which most match area AUD, according to a pointwise method17. The
2.36 bottom row displays the 3 genes which most match AUD according to a method which considers local geometry18 The
2.37 pointwise method in the top row identifies genes which express more strongly in AUD than outside of it; its weakness is
2.38 that this includes many areas which don’t have a salient border matching the areal border. The geometric method identifies
2.39 @@ -403,13 +403,12 @@
2.40 genes which don’t express over the entire area. Genes which have high rankings using both pointwise and border criteria,
2.41 such as Aph1a in the example, may be particularly good markers. None of these genes are, individually, a perfect marker
2.42 for AUD; we deliberately chose a “difficult” area in order to better contrast pointwise with geometric methods.
2.43 -Combinations of multiple genes are useful
2.44 -_________________________________________
2.45 +Combinations of multiple genes are useful and necessary for some areas
2.46 +In Figure 3, we give an example of a cortical area which is not marked by any single gene, but which can be identified
2.47 +combinatorially.
2.48 +____________________________
2.49 17For each gene, a logistic regression in which the response variable was whether or not a surface pixel was within area AUD, and the predictor
2.50 variable was the value of the expression of the gene underneath that pixel. The resulting scores were used to rank the genes in terms of how well
2.51 -they predict area AUD.
2.52 - 18For each gene the gradient similarity between (a) a map of the expression of each gene on the cortical surface and (b) the shape of area AUD,
2.53 -was calculated, and this was used to rank the genes.
2.54
2.55
2.56
2.57 @@ -418,38 +417,25 @@
2.58 left to right and top to bottom, the genes are Ssr1, Efcbp1, Aph1a, Ptk7, Aph1a again, and Lepr
2.59
2.60
2.61 -Figure 3: Upper left: wwc1. Upper right: mtif2. Lower left: wwc1 + mtif2 (each pixel’s value on the lower left is the sum
2.62 -of the corresponding pixels in the upper row).
2.63 -Here we give an example of a cortical area which is not marked by any single gene, but which can be identified combi-
2.64 -natorially. according to logistic regression, gene wwc119 is the best fit single gene for predicting whether or not a pixel on
2.65 -the cortical surface belongs to the motor area (area MO). The upper-left picture in Figure shows wwc1’s spatial expression
2.66 -pattern over the cortex. The lower-right boundary of MO is represented reasonably well by this gene, however the gene
2.67 -overshoots the upper-left boundary. This flattened 2-D representation does not show it, but the area corresponding to the
2.68 -overshoot is the medial surface of the cortex. MO is only found on the lateral surface.
2.69 -Gene mtif220 is shown in figure the upper-right of Fig. . Mtif2 captures MO’s upper-left boundary, but not its lower-right
2.70 -boundary. Mtif2 does not express very much on the medial surface. By adding together the values at each pixel in these
2.71 -two figures, we get the lower-left of Figure . This combination captures area MO much better than any single gene.
2.72 +Figure 3: Upper left: wwc1. Upper right: mtif2. Lower left: wwc1 + mtif2 (each pixel’s value on the lower left is the
2.73 +sum of the corresponding pixels in the upper row). Acccording to logistic regression, gene wwc1 is the best fit single gene
2.74 +for predicting whether or not a pixel on the cortical surface belongs to the motor area (area MO). The upper-left picture in
2.75 +Figure 3 shows wwc1’s spatial expression pattern over the cortex. The lower-right boundary of MO is represented reasonably
2.76 +well by this gene, however the gene overshoots the upper-left boundary. This flattened 2-D representation does not show
2.77 +it, but the area corresponding to the overshoot is the medial surface of the cortex. MO is only found on the lateral surface.
2.78 +Gene mtif2 is shown in the upper-right. Mtif2 captures MO’s upper-left boundary, but not its lower-right boundary. Mtif2
2.79 +does not express very much on the medial surface. By adding together the values at each pixel in these two figures, we get
2.80 +the lower-left image. This combination captures area MO much better than any single gene.
2.81 +
2.82 +
2.83 + Figure 4: Gene Pitx2 is selectively underexpressed in area SS (somatosensory).
2.84 Underexpression of a gene can serve as a marker Underexpression of a gene can sometimes serve as a marker.
2.85 -See, for example, Figure .
2.86 +See, for example, Figure 4.
2.87 Specific to Aim 1 (and Aim 3)
2.88 Forward stepwise logistic regression todo
2.89 SVM on all genes at once
2.90 In order to see how well one can do when looking at all genes at once, we ran a support vector machine to classify cortical
2.91 -_________________________________________
2.92 - 19“WW, C2 and coiled-coil domain containing 1”; EntrezGene ID 211652
2.93 - 20“mitochondrial translational initiation factor 2”; EntrezGene ID 76784
2.94 -
2.95 -
2.96 - Figure 4: Gene Pitx2 is selectively underexpressed in area SS (somatosensory).
2.97 -
2.98 -
2.99 -Figure 5: From left to right and top to bottom, single genes which roughly identify areas SS (somatosensory primary +
2.100 -supplemental), SSs (supplemental somatosensory), PIR (piriform), FRP (frontal pole), RSP (retrosplenial), COApm (Corti-
2.101 -cal amygdalar, posterior part, medial zone). Grouping some areas together, we have also found genes to identify the groups
2.102 -ACA+PL+ILA+DP+ORB+MO (anterior cingulate, prelimbic, infralimbic, dorsal peduncular, orbital, motor), posterior
2.103 -and lateral visual (VISpm, VISpl, VISI, VISp; posteromedial, posterolateral, lateral, and primary visual). The genes are
2.104 -Pitx2, Aldh1a2, Ppfibp1, Slco1a5, Tshz2, Trhr, Col12a1, Ets1.
2.105 -surface pixels based on their gene expression profiles. We achieved classification accuracy of about 81%21. As noted above,
2.106 +surface pixels based on their gene expression profiles. We achieved classification accuracy of about 81%19. As noted above,
2.107 however, a classifier that looks at all the genes at once isn’t practically useful.
2.108 The requirement to find combinations of only a small number of genes limits us from straightforwardly applying many
2.109 of the most simple techniques from the field of supervised machine learning. In the parlance of machine learning, our task
2.110 @@ -458,7 +444,7 @@
2.111 todo
2.112 Areas which can be identified by single genes
2.113 Using all of the methods we have tried to far, we have already found single genes which roughly identify some areas and
2.114 -groupings of areas. For each of these areas, an example of a gene which roughly identifies it is shown in Figure . We have
2.115 +groupings of areas. For each of these areas, an example of a gene which roughly identifies it is shown in Figure 5. We have
2.116 not yet cross-verified these genes in other atlases.
2.117 In addition, there are a number of areas which are almost identified by single genes: COAa+NLOT (anterior part of
2.118 cortical amygdalar area, nucleus of the lateral olfactory tract), ENT (entorhinal), ACAv (ventral anterior cingulate), VIS
2.119 @@ -470,9 +456,21 @@
2.120 Dimensionality reduction plus K-means or spectral clustering
2.121 Many areas are captured by clusters of genes
2.122 todo
2.123 +todo
2.124 _________________________________________
2.125 - 215-fold cross-validation.
2.126 -todo
2.127 +they predict area AUD.
2.128 + 18For each gene the gradient similarity between (a) a map of the expression of each gene on the cortical surface and (b) the shape of area AUD,
2.129 +was calculated, and this was used to rank the genes.
2.130 + 195-fold cross-validation.
2.131 +
2.132 +
2.133 +
2.134 +Figure 5: From left to right and top to bottom, single genes which roughly identify areas SS (somatosensory primary +
2.135 +supplemental), SSs (supplemental somatosensory), PIR (piriform), FRP (frontal pole), RSP (retrosplenial), COApm (Corti-
2.136 +cal amygdalar, posterior part, medial zone). Grouping some areas together, we have also found genes to identify the groups
2.137 +ACA+PL+ILA+DP+ORB+MO (anterior cingulate, prelimbic, infralimbic, dorsal peduncular, orbital, motor), posterior
2.138 +and lateral visual (VISpm, VISpl, VISI, VISp; posteromedial, posterolateral, lateral, and primary visual). The genes are
2.139 +Pitx2, Aldh1a2, Ppfibp1, Slco1a5, Tshz2, Trhr, Col12a1, Ets1.
2.140 Research plan
2.141 Further work on flatmapping
2.142 In anatomy, the manifold of interest is usually either defined by a combination of two relevant anatomical axes (todo),
3.1 Binary file grant.odt has changed
4.1 Binary file grant.pdf has changed
5.1 --- a/grant.txt Sun Apr 19 14:29:24 2009 -0700
5.2 +++ b/grant.txt Sun Apr 19 14:44:41 2009 -0700
5.3 @@ -83,15 +83,17 @@
5.4 Atlas". AGEA has three
5.5 components:
5.6
5.7 -* Gene Finder: The user selects a seed voxel and the system (1) chooses a
5.8 +\begin{itemize}
5.9 +\item Gene Finder: The user selects a seed voxel and the system (1) chooses a
5.10 cluster which includes the seed voxel, (2) yields a list of genes
5.11 which are overexpressed in that cluster. (note: the ABA website also contains pre-prepared lists of overexpressed genes for selected structures)
5.12
5.13 -* Correlation: The user selects a seed voxel and
5.14 +\item Correlation: The user selects a seed voxel and
5.15 the shows the user how much correlation there is between the gene
5.16 expression profile of the seed voxel and every other voxel.
5.17
5.18 -* Clusters: will be described later
5.19 +\item Clusters: will be described later
5.20 +\end{itemize}
5.21
5.22 Gene Finder is different from our Aim 1 in at least three ways. First, Gene Finder finds only single genes, whereas we will also look for combinations of genes. Second, gene finder can only use overexpression as a marker, whereas we will also search for underexpression. Third, Gene Finder uses a simple pointwise score\footnote{"Expression energy ratio", which captures overexpression.}, whereas we will also use geometric scores such as gradient similarity. The Preliminary Data section contains evidence that each of our three choices is the right one.
5.23
5.24 @@ -265,7 +267,7 @@
5.25
5.26 One of the simplest methods in this class is to use correlation as the match score. We calculated the correlation between each gene and each cortical area. The top row of Figure \ref{SScorrLr} shows the three genes most correlated with area SS.
5.27
5.28 -\begin{figure}\label{SScorrLr}
5.29 +\begin{figure}
5.30 \includegraphics[scale=.31]{singlegene_SS_corr_top_1_2365_jet.eps}
5.31 \includegraphics[scale=.31]{singlegene_SS_corr_top_2_242_jet.eps}
5.32 \includegraphics[scale=.31]{singlegene_SS_corr_top_3_654_jet.eps}\\
5.33 @@ -276,7 +278,7 @@
5.34
5.35
5.36 \caption{Top row: Genes Nfic, A930001M12Rik, C130038G02Rik are the most correlated with area SS (somatosensory cortex). Bottom row: Genes C130038G02Rik, Cacna1i, Car10 are those with the best fit using logistic regression. Within each picture, the vertical axis roughly corresponds to anterior at the top and posterior at the bottom, and the horizontal axis roughly corresponds to medial at the left and lateral at the right. The red outline is the boundary of region MO. Pixels are colored according to correlation, with red meaning high correlation and blue meaning low.}
5.37 -\end{figure}
5.38 +\label{SScorrLr}\end{figure}
5.39
5.40
5.41
5.42 @@ -308,7 +310,7 @@
5.43 To show that gradient similarity can provide useful information that cannot be detected via pointwise analyses, consider Fig. \ref{AUDgeometry}. The top row of Fig. \ref{AUDgeometry} displays the 3 genes which most match area AUD, according to a pointwise method\footnote{For each gene, a logistic regression in which the response variable was whether or not a surface pixel was within area AUD, and the predictor variable was the value of the expression of the gene underneath that pixel. The resulting scores were used to rank the genes in terms of how well they predict area AUD.}. The bottom row displays the 3 genes which most match AUD according to a method which considers local geometry\footnote{For each gene the gradient similarity between (a) a map of the expression of each gene on the cortical surface and (b) the shape of area AUD, was calculated, and this was used to rank the genes.} The pointwise method in the top row identifies genes which express more strongly in AUD than outside of it; its weakness is that this includes many areas which don't have a salient border matching the areal border. The geometric method identifies genes whose salient expression border seems to partially line up with the border of AUD; its weakness is that this includes genes which don't express over the entire area. Genes which have high rankings using both pointwise and border criteria, such as $Aph1a$ in the example, may be particularly good markers. None of these genes are, individually, a perfect marker for AUD; we deliberately chose a "difficult" area in order to better contrast pointwise with geometric methods.
5.44
5.45
5.46 -\begin{figure}\label{AUDgeometry}
5.47 +\begin{figure}
5.48 \includegraphics[scale=.31]{singlegene_AUD_lr_top_1_3386_jet.eps}
5.49 \includegraphics[scale=.31]{singlegene_AUD_lr_top_2_1258_jet.eps}
5.50 \includegraphics[scale=.31]{singlegene_AUD_lr_top_3_420_jet.eps}
5.51 @@ -317,22 +319,27 @@
5.52 \includegraphics[scale=.31]{singlegene_AUD_gr_top_2_420_jet.eps}
5.53 \includegraphics[scale=.31]{singlegene_AUD_gr_top_3_2072_jet.eps}
5.54 \caption{The top row shows the three genes which (individually) best predict area AUD, according to logistic regression. The bottom row shows the three genes which (individually) best match area AUD, according to gradient similarity. From left to right and top to bottom, the genes are $Ssr1$, $Efcbp1$, $Aph1a$, $Ptk7$, $Aph1a$ again, and $Lepr$}
5.55 -\end{figure}
5.56 -
5.57 -
5.58 -\vspace{0.3cm}**Combinations of multiple genes are useful**
5.59 -
5.60 -Here we give an example of a cortical area which is not marked by any single gene, but which can be identified combinatorially. according to logistic regression, gene wwc1\footnote{"WW, C2 and coiled-coil domain containing 1"; EntrezGene ID 211652} is the best fit single gene for predicting whether or not a pixel on the cortical surface belongs to the motor area (area MO). The upper-left picture in Figure \ref{MOcombo} shows wwc1's spatial expression pattern over the cortex. The lower-right boundary of MO is represented reasonably well by this gene, however the gene overshoots the upper-left boundary. This flattened 2-D representation does not show it, but the area corresponding to the overshoot is the medial surface of the cortex. MO is only found on the lateral surface.
5.61 -
5.62 -Gene mtif2\footnote{"mitochondrial translational initiation factor 2"; EntrezGene ID 76784} is shown in figure the upper-right of Fig. \ref{MOcombo}. Mtif2 captures MO's upper-left boundary, but not its lower-right boundary. Mtif2 does not express very much on the medial surface. By adding together the values at each pixel in these two figures, we get the lower-left of Figure \ref{MOcombo}. This combination captures area MO much better than any single gene.
5.63 -
5.64 -\begin{figure}\label{MOcombo}
5.65 +\label{AUDgeometry}\end{figure}
5.66 +
5.67 +
5.68 +\vspace{0.3cm}**Combinations of multiple genes are useful and necessary for some areas**
5.69 +
5.70 +In Figure \ref{MOcombo}, we give an example of a cortical area which is not marked by any single gene, but which can be identified combinatorially.
5.71 +
5.72 +%% wwc1\footnote{"WW, C2 and coiled-coil domain containing 1"; EntrezGene ID 211652}
5.73 +%% mtif2\footnote{"mitochondrial translational initiation factor 2"; EntrezGene ID 76784}
5.74 +
5.75 +%%Acccording to logistic regression, gene wwc1\footnote{"WW, C2 and coiled-coil domain containing 1"; EntrezGene ID 211652} is the best fit single gene for predicting whether or not a pixel on the cortical surface belongs to the motor area (area MO). The upper-left picture in Figure \ref{MOcombo} shows wwc1's spatial expression pattern over the cortex. The lower-right boundary of MO is represented reasonably well by this gene, however the gene overshoots the upper-left boundary. This flattened 2-D representation does not show it, but the area corresponding to the overshoot is the medial surface of the cortex. MO is only found on the lateral surface.
5.76 +
5.77 +%%Gene mtif2\footnote{"mitochondrial translational initiation factor 2"; EntrezGene ID 76784} is shown in figure the upper-right of Fig. \ref{MOcombo}. Mtif2 captures MO's upper-left boundary, but not its lower-right boundary. Mtif2 does not express very much on the medial surface. By adding together the values at each pixel in these two figures, we get the lower-left of Figure \ref{MOcombo}. This combination captures area MO much better than any single gene.
5.78 +
5.79 +\begin{figure}
5.80 \includegraphics[scale=.36]{MO_vs_Wwc1_jet.eps}
5.81 \includegraphics[scale=.36]{MO_vs_Mtif2_jet.eps}
5.82
5.83 \includegraphics[scale=.36]{MO_vs_Wwc1_plus_Mtif2_jet.eps}
5.84 -\caption{Upper left: $wwc1$. Upper right: $mtif2$. Lower left: wwc1 + mtif2 (each pixel's value on the lower left is the sum of the corresponding pixels in the upper row). }
5.85 -\end{figure}
5.86 +\caption{Upper left: $wwc1$. Upper right: $mtif2$. Lower left: wwc1 + mtif2 (each pixel's value on the lower left is the sum of the corresponding pixels in the upper row). Acccording to logistic regression, gene wwc1 is the best fit single gene for predicting whether or not a pixel on the cortical surface belongs to the motor area (area MO). The upper-left picture in Figure \ref{MOcombo} shows wwc1's spatial expression pattern over the cortex. The lower-right boundary of MO is represented reasonably well by this gene, however the gene overshoots the upper-left boundary. This flattened 2-D representation does not show it, but the area corresponding to the overshoot is the medial surface of the cortex. MO is only found on the lateral surface. Gene mtif2 is shown in the upper-right. Mtif2 captures MO's upper-left boundary, but not its lower-right boundary. Mtif2 does not express very much on the medial surface. By adding together the values at each pixel in these two figures, we get the lower-left image. This combination captures area MO much better than any single gene. }
5.87 +\label{MOcombo}\end{figure}
5.88
5.89
5.90
5.91 @@ -343,10 +350,10 @@
5.92 Underexpression of a gene can sometimes serve as a marker. See, for example, Figure \ref{hole}.
5.93
5.94
5.95 -\begin{figure}\label{hole}
5.96 +\begin{figure}
5.97 \includegraphics[scale=.31]{holeExample_2682_SS_jet.eps}
5.98 \caption{Gene Pitx2 is selectively underexpressed in area SS (somatosensory).}
5.99 -\end{figure}
5.100 +\label{hole}\end{figure}
5.101
5.102
5.103
5.104 @@ -374,7 +381,7 @@
5.105 In addition, there are a number of areas which are almost identified by single genes: COAa+NLOT (anterior part of cortical amygdalar area, nucleus of the lateral olfactory tract), ENT (entorhinal), ACAv (ventral anterior cingulate), VIS (visual), AUD (auditory).
5.106
5.107
5.108 -\begin{figure}\label{singleSoFar}
5.109 +\begin{figure}
5.110 \includegraphics[scale=.31]{singlegene_example_2682_Pitx2_SS_jet.eps}
5.111 \includegraphics[scale=.31]{singlegene_example_371_Aldh1a2_SSs_jet.eps}
5.112 \includegraphics[scale=.31]{singlegene_example_2759_Ppfibp1_PIR_jet.eps}
5.113 @@ -385,7 +392,7 @@
5.114 \includegraphics[scale=.31]{singlegene_example_1334_Ets1_post_lat_vis_jet.eps}
5.115
5.116 \caption{From left to right and top to bottom, single genes which roughly identify areas SS (somatosensory primary + supplemental), SSs (supplemental somatosensory), PIR (piriform), FRP (frontal pole), RSP (retrosplenial), COApm (Cortical amygdalar, posterior part, medial zone). Grouping some areas together, we have also found genes to identify the groups ACA+PL+ILA+DP+ORB+MO (anterior cingulate, prelimbic, infralimbic, dorsal peduncular, orbital, motor), posterior and lateral visual (VISpm, VISpl, VISI, VISp; posteromedial, posterolateral, lateral, and primary visual). The genes are $Pitx2$, $Aldh1a2$, $Ppfibp1$, $Slco1a5$, $Tshz2$, $Trhr$, $Col12a1$, $Ets1$.}
5.117 -\end{figure}
5.118 +\label{singleSoFar}\end{figure}
5.119
5.120
5.121