1
Small-sample protocol recommendations, and GeneChip hybridization followed our standard protocol [28]. After washes, arrays were scanned using the GeneChip Scanner 3000 (Affymetrix, Santa Clara, CA). Light intensity data from all spots on each chip were recorded as .cel files, stored on an internal server, and written to a digital variable disc (DVD).Data normalization and filteringreproducibility
1
Cal significance.Paraffin sections of our patient-derived glioblastoma xenografts (15 of 22 lines) were stained for galectin-1 expression. Around half of the xenografts tested showed preferential staining at the tumor-brain interface (Figure 3). A few tumors stained in their entirety, and another subset lacked significant staining. The 2 to 4 fold change in galectin-1 mRNA expression at the tumor
1
S independently with similar results. Statistical significance was determined by one-way ANOVA followed by Bonferroni's post hoc analysis for multiple comparisons. *Significantly different compared with PBStreated control (P
1
Filtering algorithm. This algorithm was designed to minimize the effect of potential contamination of the edge samples with normal mouse brain cells. Relative expression values for each gene from tumor core, tumor edge, and normal mouse brain samples were compared. Genes of interest were identified that met three criteria: a) low expression at tumor core; b) relatively increased expression at tumo
1
Galectin-1 transfectants. A population of GFP-sorted cells (the "Gal-1" bars in Figure 4A) was compared to its parental counterpart. The number of metabolically-active cells attached to fibronectin was no different between the two lines at eight hours. Changing the media at four hours reduced the number of cells left for labeling, but the effect was equal in both groups, suggesting a similar rate
1
Filtering algorithm. This algorithm was designed to minimize the effect of potential contamination of the edge samples with normal mouse brain cells. Relative expression values for each gene from tumor core, tumor edge, and normal mouse brain samples were compared. Genes of interest were identified that met three criteria: a) low expression at tumor core; b) relatively increased expression at tumo
1
Ns.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.Toussaint et al. Molecular Cancer 2012, 11:32 http://www.molecular-cancer.com/content/11/1/Page 2 ofgrowth factor receptors [11] to effector metallo- [12,13] and serine- [14] proteases. Galectin-1 has also been identified as a key player in GBM cell mi
1
Galectin-1 transfectants. A population of GFP-sorted cells (the "Gal-1" bars in Figure 4A) was compared to its parental counterpart. The number of metabolically-active cells attached to fibronectin was no different between the two lines at eight hours. Changing the media at four hours reduced the number of cells left for labeling, but the effect was equal in both groups, suggesting a similar rate