cg
changeset 32:70bdcd6c9972
.
author | bshanks@bshanks.dyndns.org |
---|---|
date | Mon Apr 13 14:31:11 2009 -0700 (16 years ago) |
parents | 95910357b4ac |
children | 6d023f15572e |
files | grant.html grant.odt grant.pdf grant.txt |
line diff
1.1 --- a/grant.html Mon Apr 13 04:07:32 2009 -0700
1.2 +++ b/grant.html Mon Apr 13 14:31:11 2009 -0700
1.3 @@ -205,13 +205,28 @@
1.4 similarity, which is discussed in Preliminary Work) may be necessary in order to achieve the best results in this application.
1.5 We are aware of two existing efforts to relate spatial gene expression data to anatomy through computational methods.
1.6 [? ] describes an analysis of the anatomy of the hippocampus using the ABA dataset. In addition to manual analysis,
1.7 -two clustering methods were employed, a modified Non-negative Matrix Factorization (NNMF), and a hierarchial bifurcation
1.8 -clustering scheme based on correlation as the similarity score. The paper yielded impressive results, proving the usefulness
1.9 -of such research. We have run NNMF on the cortical dataset and while the results are promising (see Preliminary Data), we
1.10 -think that it will be possible to find a better method3 (we also think that more automation of the parts that this paper’s
1.11 -authors did manually will be possible).
1.12 -and [?] describes AGEA. todo
1.13 -_____________
1.14 +two clustering methods were employed, a modified Non-negative Matrix Factorization (NNMF), and a hierarchial recursive
1.15 +bifurcation clustering scheme based on correlation as the similarity score. The paper yielded impressive results, proving the
1.16 +usefulness of such research. We have run NNMF on the cortical dataset and while the results are promising (see Preliminary
1.17 +Data), we think that it will be possible to find a better method3 (we also think that more automation of the parts that this
1.18 +paper’s authors did manually will be possible).
1.19 +[? ] describes AGEA, ”Anatomic Gene Expression Atlas”. AGEA is an analysis tool for the ABA dataset. AGEA has
1.20 +three components:
1.21 +* Gene Finder: The user selects a seed voxel and the system (1) chooses a cluster which includes the seed voxel, (2) yields
1.22 +a list of genes which are overexpressed in that cluster.
1.23 +* Correlation: The user selects a seed voxel and the shows the user how much correlation there is between the gene
1.24 +expression profile of the seed voxel and every other voxel.
1.25 +* Clusters: AGEA includes a precomputed hierarchial clustering of voxels based on a recursive bifurcation algorithm with
1.26 +correlation as the similarity metric.
1.27 +At first glance AGEA seems similar to this proposal, but in fact it is different.
1.28 +Gene Finder is different from our Aim 1 in at least four ways. First, although the user chooses a seed voxel, Gene Finder,
1.29 +not the user, chooses the cluster for which genes will be found, and in our experience it never chooses cortical areas, instead
1.30 +preferring cortical layers. Therefore, Gene Finder cannot be used to find marker genes for cortical areas. Second, Gene Finder
1.31 +finds only single genes, whereas we will also look for combinations of genes. Third, gene finder can only use overexpression
1.32 +as a marker, whereas we will also look for underexpression. Fourth, Gene Finder uses a simple pointwise metric (“expression
1.33 +energy ratio”, which captures overexpression), whereas we will also use geometric metrics such as gradient similarity.
1.34 +The hierarchial clustering is different from our Aim 2 in at least two ways. todo
1.35 +_________________________________________
1.36 3We ran “vanilla” NNMF, whereas the paper under discussion used a modified method. Their main modification consisted of adding a soft
1.37 spatial contiguity constraint. However, on our dataset, NNMF naturally produced spatially contiguous clusters, so no additional constraint was
1.38 needed. The paper under discussion mentions that they also tried a hierarchial variant of NNMF, but since they didn’t report its results, we assume
2.1 Binary file grant.odt has changed
3.1 Binary file grant.pdf has changed
4.1 --- a/grant.txt Mon Apr 13 04:07:32 2009 -0700
4.2 +++ b/grant.txt Mon Apr 13 14:31:11 2009 -0700
4.3 @@ -144,10 +144,31 @@
4.4
4.5 We are aware of two existing efforts to relate spatial gene expression data to anatomy through computational methods.
4.6
4.7 -\cite{thompson_genomic_2008} describes an analysis of the anatomy of the hippocampus using the ABA dataset. In addition to manual analysis, two clustering methods were employed, a modified Non-negative Matrix Factorization (NNMF), and a hierarchial bifurcation clustering scheme based on correlation as the similarity score. The paper yielded impressive results, proving the usefulness of such research. We have run NNMF on the cortical dataset and while the results are promising (see Preliminary Data), we think that it will be possible to find a better method\footnote{We ran "vanilla" NNMF, whereas the paper under discussion used a modified method. Their main modification consisted of adding a soft spatial contiguity constraint. However, on our dataset, NNMF naturally produced spatially contiguous clusters, so no additional constraint was needed. The paper under discussion mentions that they also tried a hierarchial variant of NNMF, but since they didn't report its results, we assume that those result were not any more impressive than the results of the non-hierarchial variant.} (we also think that more automation of the parts that this paper's authors did manually will be possible).
4.8 -
4.9 -
4.10 - and \cite{ng_anatomic_2009} describes AGEA. todo
4.11 +\cite{thompson_genomic_2008} describes an analysis of the anatomy of
4.12 +the hippocampus using the ABA dataset. In addition to manual analysis,
4.13 +two clustering methods were employed, a modified Non-negative Matrix
4.14 +Factorization (NNMF), and a hierarchial recursive bifurcation clustering scheme based on correlation as the similarity score. The paper yielded impressive results, proving the usefulness of such research. We have run NNMF on the cortical dataset and while the results are promising (see Preliminary Data), we think that it will be possible to find a better method\footnote{We ran "vanilla" NNMF, whereas the paper under discussion used a modified method. Their main modification consisted of adding a soft spatial contiguity constraint. However, on our dataset, NNMF naturally produced spatially contiguous clusters, so no additional constraint was needed. The paper under discussion mentions that they also tried a hierarchial variant of NNMF, but since they didn't report its results, we assume that those result were not any more impressive than the results of the non-hierarchial variant.} (we also think that more automation of the parts that this paper's authors did manually will be possible).
4.15 +
4.16 +
4.17 +\cite{ng_anatomic_2009} describes AGEA, "Anatomic Gene Expression
4.18 +Atlas". AGEA is an analysis tool for the ABA dataset. AGEA has three
4.19 +components:
4.20 +
4.21 +* Gene Finder: The user selects a seed voxel and the system (1) chooses a
4.22 +cluster which includes the seed voxel, (2) yields a list of genes
4.23 +which are overexpressed in that cluster.
4.24 +
4.25 +* Correlation: The user selects a seed voxel and
4.26 +the shows the user how much correlation there is between the gene
4.27 +expression profile of the seed voxel and every other voxel.
4.28 +
4.29 +* Clusters: AGEA includes a precomputed hierarchial clustering of voxels based on a recursive bifurcation algorithm with correlation as the similarity metric.
4.30 +
4.31 +At first glance AGEA seems similar to this proposal, but in fact it is different.
4.32 +
4.33 +Gene Finder is different from our Aim 1 in at least four ways. First, although the user chooses a seed voxel, Gene Finder, not the user, chooses the cluster for which genes will be found, and in our experience it never chooses cortical areas, instead preferring cortical layers. Therefore, Gene Finder cannot be used to find marker genes for cortical areas. Second, Gene Finder finds only single genes, whereas we will also look for combinations of genes. Third, gene finder can only use overexpression as a marker, whereas we will also look for underexpression. Fourth, Gene Finder uses a simple pointwise metric ("expression energy ratio", which captures overexpression), whereas we will also use geometric metrics such as gradient similarity.
4.34 +
4.35 +The hierarchial clustering is different from our Aim 2 in at least two ways. todo
4.36
4.37
4.38