Gene selection (feature selection) is generally pertormed in gene space(feature space), where a very serious curse of dimensionality problem always existsbecause the number of genes is much larger than the number of s...Gene selection (feature selection) is generally pertormed in gene space(feature space), where a very serious curse of dimensionality problem always existsbecause the number of genes is much larger than the number of samples in gene space(G-space). This results in difficulty in modeling the data set in this space and the lowconfidence of the result of gene selection. How to find a gene subset in this case is achallenging subject. In this paper, the above G-space is transformed into its dual space,referred to as class space (C-space) such that the number of dimensions is the verynumber of classes of the samples in G-space and the number of samples in C-space isthe number of genes in G-space. it is obvious that the curse of dimensionality in C-spacedoes not exist. A new gene selection method which is based on the principle of separatingdifferent classes as far as possible is presented with the help of Principal ComponentAnalysis (PCA). The experimental results on gene selection for real data set areevaluated with Fisher criterion, weighted Fisher criterion as well as leave-one-out crossvalidation, showing that the method presented here is effective and efficient.展开更多
Retina is a multilayer and highly specialized tissue important in converting light into neural signals. In humans, the critical period for the formation of complex multiplayer structure takes place during embryogenesi...Retina is a multilayer and highly specialized tissue important in converting light into neural signals. In humans, the critical period for the formation of complex multiplayer structure takes place during embryogenesis be- tween 12 and 28 weeks. The morphologic changes during retinal development in humans have been studied but little is known about the molecular events essential for the formation of the retina. To gain further insights into this process, cDNA microarrays containing 16361 human gene probes were used to measure the gene expression levels in retinas. Of the 16361 genes, 68.7%, 71.4% and 69.7% showed positive hybridiza- tion with cDNAs made from 12—16 week fetal, 22—26 week fetal and adult retinas. A total of 814 genes showed a mini- mum of 3-fold changes between the lowest and highest ex- pression levels among three time points and among them, 106 genes had expression levels with the hybridization intensity above 100 at one or more time points. The clustering analysis suggested that the majority of differentially expressed genes were down-regulated during the retinal development. The differentially expressed genes were further classified accord- ing to functions of known genes, and were ranked in de- creasing order according to frequency: development, differ- entiation, signal transduction, protein synthesis and transla- tion, metabolism, DNA binding and transcription, DNA syn- thesis-repair-recombination, immuno-response, ion channel- transport, cell receptor, cytoskeleton, cell cycle, pro-oncogene, stress and apoptosis related genes. Among these 106 differen- tially expressed genes, 60 are already present in NEI retina cDNA or EST Databank but the remaining 46 genes are ab- sent and thus identified as “function unknown”. To validate gene expression data from the microarray, real-time RT-PCR was performed for 46 “function unknown” genes and 6 known retina specific expression genes, and β-actin was used as internal control. Twenty-seven of these genes showed very similar expression profiles between the microarray and real-time RT-PCR data. In situ hybridization revealed both expression level and cellular distribution of NNAT in retina. Finally, the chromosomal locations of 106 differentially ex- pressed genes were also searched and one of these genes is associated with autosomal dominant cone or cone-rod dys- trophy. The data from present study provide insights into understanding genetic programs during human retinal de- velopment and help identify additional retinal disease genes.展开更多
文摘Gene selection (feature selection) is generally pertormed in gene space(feature space), where a very serious curse of dimensionality problem always existsbecause the number of genes is much larger than the number of samples in gene space(G-space). This results in difficulty in modeling the data set in this space and the lowconfidence of the result of gene selection. How to find a gene subset in this case is achallenging subject. In this paper, the above G-space is transformed into its dual space,referred to as class space (C-space) such that the number of dimensions is the verynumber of classes of the samples in G-space and the number of samples in C-space isthe number of genes in G-space. it is obvious that the curse of dimensionality in C-spacedoes not exist. A new gene selection method which is based on the principle of separatingdifferent classes as far as possible is presented with the help of Principal ComponentAnalysis (PCA). The experimental results on gene selection for real data set areevaluated with Fisher criterion, weighted Fisher criterion as well as leave-one-out crossvalidation, showing that the method presented here is effective and efficient.
文摘Retina is a multilayer and highly specialized tissue important in converting light into neural signals. In humans, the critical period for the formation of complex multiplayer structure takes place during embryogenesis be- tween 12 and 28 weeks. The morphologic changes during retinal development in humans have been studied but little is known about the molecular events essential for the formation of the retina. To gain further insights into this process, cDNA microarrays containing 16361 human gene probes were used to measure the gene expression levels in retinas. Of the 16361 genes, 68.7%, 71.4% and 69.7% showed positive hybridiza- tion with cDNAs made from 12—16 week fetal, 22—26 week fetal and adult retinas. A total of 814 genes showed a mini- mum of 3-fold changes between the lowest and highest ex- pression levels among three time points and among them, 106 genes had expression levels with the hybridization intensity above 100 at one or more time points. The clustering analysis suggested that the majority of differentially expressed genes were down-regulated during the retinal development. The differentially expressed genes were further classified accord- ing to functions of known genes, and were ranked in de- creasing order according to frequency: development, differ- entiation, signal transduction, protein synthesis and transla- tion, metabolism, DNA binding and transcription, DNA syn- thesis-repair-recombination, immuno-response, ion channel- transport, cell receptor, cytoskeleton, cell cycle, pro-oncogene, stress and apoptosis related genes. Among these 106 differen- tially expressed genes, 60 are already present in NEI retina cDNA or EST Databank but the remaining 46 genes are ab- sent and thus identified as “function unknown”. To validate gene expression data from the microarray, real-time RT-PCR was performed for 46 “function unknown” genes and 6 known retina specific expression genes, and β-actin was used as internal control. Twenty-seven of these genes showed very similar expression profiles between the microarray and real-time RT-PCR data. In situ hybridization revealed both expression level and cellular distribution of NNAT in retina. Finally, the chromosomal locations of 106 differentially ex- pressed genes were also searched and one of these genes is associated with autosomal dominant cone or cone-rod dys- trophy. The data from present study provide insights into understanding genetic programs during human retinal de- velopment and help identify additional retinal disease genes.
