PLSR model loadings (pMEK and pERK included) from analysis of the covariation of molecular signals with cell viability and apoptosis fraction in BRAF(V600E/D) melanoma cell lines - Dataset (ID:20229)
|HMS Dataset ID:||20229|
|Dataset Title:||PLSR model loadings (pMEK and pERK included) from analysis of the covariation of molecular signals with cell viability and apoptosis fraction in BRAF(V600E/D) melanoma cell lines|
|Publication(s) Using Dataset:||PMID: 25814555|
|Project Summary Page(s):||lincs.hms.harvard.edu/fallahi-sichani-molsystbiol-2015|
|Screening Lab Investigator:||Mohammad Fallahi-Sichani|
|Screening Principal Investigator:||Peter Sorger|
|Assay Description:||This analysis dataset presents partial-least-squares regression (PLSR) model loadings arising from analysis of HMS LINCS datasets #20217 and 20218 for covariation of molecular signals (measured by RPPA; dataset #20218) with cellular responses (relative viability and apoptotic fractions; datasets #20217) when pMEK and pERK signals are included.|
1. Cellular responses at the level of relative viability, apoptotic fraction, and protein phosphorylation were assessed as described in datasets #20217 and 20218.|
2. Partial-least-squares regression (PLSR) modeling (Geladi & Kowalski, 1986; Janes & Yaffe, 2006) was used to identify statistically significant covariation between molecular signals (input data; measured by RPPA) and corresponding cellular responses (output data; relative viability and apoptotic fractions) for each cell line, following the specifications given below.
3. All RPPA measurements (dataset #20218) were used, and thus the dimensions of the input data matrix for each cell line were 35×105 (5 drugs × 7 doses; 21 signals × 5 time points).
4. The initial dimensions for the cellular response measurements (dataset #20217) were 35×6 (5 drugs × 7 doses; viability and apoptotic fraction at 3 time points). We combined the two cellular response measurements (viability and apoptosis) at different time points to generate a new variable, “non-apoptotic viability”, by subtracting the number of apoptotic cells from the total number of cells at each condition followed by normalization to a DMSO-treated control. We then averaged the 48 hr and 72 hr non-apoptotic viability data to generate one output variable for each of the 35 conditions in the PLSR model. By averaging the cellular responses across the two time points, we account for the rate at which different cell lines respond to treatment. (We did not use the 24 hr cellular response data for PLSR modeling, as most of the cell lines do not begin to respond to treatments in the first 24 hr.) Accounting for averaging, the dimensions of the output data used in the PLSR models was 35 × 1. In the case of one cell line (K2), cellular response data at 72 hr were unavailable, so we used the 48 hr data for PLSR modeling.
5. All data were mean centered and unit variance scaled (z-score scaled) across all conditions and time points.
6. PLSR analysis was performed using MATLAB R2012b and the “plsregress” function (using two PLSR components). PLSR component 1 and 2 (PC1 and PC2) loadings are reported.
|Assay Protocol Reference:||
Geladi P, Kowalski BR (1986) Partial least-squares regression: a tutorial. Anal Chim Acta 185:1-17.|
Janes KA, Yaffe MB (2006) Data-driven modelling of signal-transduction networks. Nat Rev Mol Cell Biol 7:820-828.
|HMS Dataset Type:||Analysis|
|Date Publicly Available:||2015-04-17|
|Most Recent Update:||2015-09-23|