Department of Cell and Molecular Biology-Genetics, Göteborg University, Box 462, SE 40530 Göteborg, Sweden.
BMC Genomics. 2010 Mar 2;11:146. doi: 10.1186/1471-2164-11-146.
Rat models are frequently used to link genomic regions to experimentally induced arthritis in quantitative trait locus (QTL) analyses. To facilitate the search for candidate genes within such regions, we have previously developed an application (CGC) that uses weighted keywords to rank genes based on their descriptive text. In this study, CGC is used for analyzing the localization of candidate genes from two viewpoints: distribution over the rat genome and functional connections between arthritis QTLs.
To investigate if candidate genes identified by CGC are more likely to be found inside QTLs, we ranked 2403 genes genome wide in rat. The number of genes within different ranges of CGC scores localized inside and outside QTLs was then calculated. Furthermore, we investigated if candidate genes within certain QTLs share similar functions, and if these functions could be connected to genes within other QTLs. Based on references between genes in OMIM, we created connections between genes in QTLs identified in two distinct rat crosses. In this way, QTL pairs with one QTL from each cross that share an unexpectedly high number of gene connections were identified. The genes that were found to connect a pair of QTLs were then functionally analysed using a publicly available classification tool.
Out of the 2403 genes ranked by the CGC application, 1160 were localized within QTL regions. No difference was observed between highly and lowly rated genes. Hence, highly rated candidate genes for arthritis seem to be distributed randomly inside and outside QTLs. Furthermore, we found five pairs of QTLs that shared a significantly high number of interconnected genes. When functionally analyzed, most genes connecting two QTLs could be included in a single functional cluster. Thus, the functional connections between these genes could very well be involved in the development of an arthritis phenotype.
From the genome wide CGC search, we conclude that candidate genes for arthritis in rat are randomly distributed between QTL and non-QTL regions. We do however find certain pairs of QTLs that share a large number of functionally connected candidate genes, suggesting that these QTLs contain a number of genes involved in similar functions contributing to the arthritis phenotype.
在定量性状基因座(QTL)分析中,大鼠模型常用于将基因组区域与实验性诱导关节炎联系起来。为了方便在这些区域内寻找候选基因,我们之前开发了一种应用程序(CGC),该程序使用加权关键词根据描述性文本对基因进行排名。在这项研究中,CGC 用于从两个角度分析候选基因的定位:在大鼠基因组中的分布和关节炎 QTL 之间的功能连接。
为了研究通过 CGC 鉴定的候选基因是否更有可能位于 QTL 内,我们在大鼠中对 2403 个基因进行了全基因组排名。然后计算了位于不同 CGC 分数范围内的基因在 QTL 内和外的数量。此外,我们还研究了某些 QTL 内的候选基因是否具有相似的功能,以及这些功能是否可以与其他 QTL 内的基因连接。基于 OMIM 中基因之间的参考文献,我们在两个不同大鼠杂交中鉴定的 QTL 之间创建了基因连接。通过这种方式,鉴定出来自每个杂交的一个 QTL 的一对 QTL 之间具有异常高数量的基因连接。然后使用公共分类工具对连接一对 QTL 的基因进行功能分析。
在 CGC 应用程序排名的 2403 个基因中,有 1160 个基因定位于 QTL 区域内。高评分和低评分基因之间没有观察到差异。因此,关节炎的高评分候选基因似乎随机分布在 QTL 内和外。此外,我们发现了五对 QTL,它们共享大量相互连接的基因。当进行功能分析时,连接两个 QTL 的大多数基因可以归入单个功能群。因此,这些基因之间的功能连接很可能与关节炎表型的发展有关。
从全基因组 CGC 搜索中,我们得出结论,大鼠关节炎的候选基因在 QTL 和非 QTL 区域之间随机分布。然而,我们确实发现了某些 QTL 对,它们共享大量功能上相互连接的候选基因,这表明这些 QTL 包含了许多参与相似功能的基因,这些基因共同导致了关节炎表型。