자유게시판

• 목록
• 답변
• 글쓰기
• 게시판 리스트 옵션
  • 수정
  • 삭제

Recent evidence suggests the miR-29 family consisting
S in this pathway. Recent evidence suggests the miR-29 family consisting of miR-29a, miR-29b and miR-29c are antiangiogenic [50]. Both approaches recovered miR-29 targeting of beta integrin subunit (ITGB1) in the active subnetwork. The Original Angiogenesis Network also contained an interaction from the GeneGO database showing miR-29b targeting of collagen IV gene COL4A1 [51,52]. This interaction was not recovered by either the single parameter or full parameter approach. We might have thought the full parameter approach would detect this interaction, since miR-29b is already considered to be active in its role in inhibiting the pro-angiogenic integrins. However, in this situation, there was no evidence of gene expression of the collagen genes in our input gene list, and thus possibly the low artificial correlation score (the noise component) was more influential than the local neighborhood influence (MRF component) in this circumstance.Figure 4 contains the interactions from GeneGO Pathway Map: Cytoskeleton Remodeling - Integrin Inside Out Signaling, overlaid with the extra interactions detected using the full parameter approach.Application to mouse organogenesis dataIn order to test our method using actual Spearman correlation values in the likelihood function, rather than our artificial scoring system used on the angiogenesis data, we applied the method to data from a developmental mouse lung organogenesis study [53], downloaded from NCBI Gene Expression Omnibus (GEO) [43], Accession GSE20954 and GSE21052. As the experiments were performed using Affymetrix chips with known 45,000 probes (mRNA data) as Methyl 6-bromo-5-fluoropicolinate opposed to the home made chips, we had expression data matching nearly all Network Objects in the Metacore 2-Bromo-4-fluoro-5-methylbenzoic acid networks, not only the so-called seed nodes. Secondly, we had a stronger signal when calculating Spearman correlation coefficients across interaction expression data. To create a set of mRNA and miRNA genes to upload to Metacore in our list we wished to identify miRNAs and mRNAs that were expected to regulate the same processes. In particular, we wanted late-onset genes that had high expression in the adulthood stage, so we could PubMed ID:https://www.ncbi.nlm.nih.gov/pubmed/13867361 compare our own biological analysis of resulting networks with the biological analysis of similar gene sets in the original network. In this work, cluster 6 mRNAs and cluster 1 miRNAs were considered to be late onset genes. For the miRNA array, we used the clusters generated by [53]Figure 4 A network showing the interactions from GeneGO pathway map: cytoskeleton remodeling - integrin inside out signaling, overlaid with the extra interactions detected using the PubMed ID:https://www.ncbi.nlm.nih.gov/pubmed/13867361 full parameter approach.Lichtenstein et al. BMC Bioinformatics 2013, 14:59 http://www.biomedcentral.com/1471-2105/14/Page 18 of(provided to us by the authors) and selected miRNA cluster 1, which also had a peak late onset adulthood among member miRNAs. To closely reproduce their results, we clustered the 11220 mRNA probes identified as active in [53]. We selected a subset using decideTests in the limma package in R in the contrast PN30.adult-PN10 as active. We then clustered the Affymetrix probe expression values over the 7 time points using hierarchical clustering in R (hclust). We created a representative expression profile for each cluster by finding the mean expression value of all probes in the cluster at each time point (see Additional file 4). Among our clusters, the representative profile for cluster 2 (containing 850 probes) showed a peak at.