6. Materials and Methods

6.1. HCT-116 Cell Lines
Two cell lines of HCT-116 (wt-p53 and null-p53) were treated with three drugs: 5-fluorouracil (5-FU), oxaliplatin (OX) and irinotecan (CPT-11). The total RNAs were isolated from the six treated cell lines and two non-treated cell lines (serving as controls) and tested with miRCURY LNA MicroRNA Array v7.5.0 (Exiqon Inc.). For HCT-116 (null-p53), the control sample was tested three times, and the CPT-11 treated sample was tested twice. For HCT-116 (wt-p53), both the control sample and the 5-FU treated sample were tested twice. A total of 359 miRNAs were profiled for each of the 13 cell lines. A total of 37 miRNAs were randomly selected and further tested using TaqMan-based qRT-PCR on an ABI 7500HT instrument (Applied Biosystems Inc., Foster City, CA, USA) [9,47,55,56,57].

6.2. Osterosarcoma Xenograft Specimens
Ten human osterosarcoma xenograft specimens were collected, and each was treated with saline (as control) and three chemotherapeutic treatments: cisplatin (CIS), doxorubicin (DOX) and ifosfamide (IFO). The total RNAs were isolated from each sample and analyzed with the following three platforms: (a) miRCURY LNA MicroRNA Array (Exiqon Inc.; Vedbaek, Denmark, based on miRbase 9.2); (b) Luminex FlexmiR MicroRNA Human Panel; and (c) TaqMan Array Human MicroRNA Panel (Applied Biosystems,Foster City, CA, USA, v2.0). A total of 577 human miRNAs were profiled with the LNA array, 391 with the Luminex bead array and 664 with the TaqMan array, where a total of 213 miRNAs were shared by all three platforms [37,40,41,58,59].

6.3. Generalized Logarithm Transformation
The following two-component measurement error model is proposed to model the measured expression levels, (1) y=α+μeη+ϵ where y is the measured raw expression level, α is the mean background noise, μ is the true expression level and η and ϵ are the multiplicative and additive measurement errors, which are assumed to be normally-distributed with mean 0 and variances ση2 and σϵ2, respectively [17,18,19]. The variance of y under this model is Var(y)=μ2Sη2+σϵ2, where Sη2=eση2(eση2−1). To ease the analyses of gene-expression microarrays using some standard statistical techniques, the following generalized logarithm transformation that stabilizes the variance has been proposed:(2) fc(z)=lnz+z2+c22 where c=σϵ/Sη. The performance of the GLOG is further studied, and simulation results show that it is a better choice compared with the “started logarithm” transformation and the “log-linear hybrid” transformation [20,60,61,62,63,64,65,66].