10/22/2021

Deposited summary table of proteins identified through

• DEP analysis of TP53 mut vs nonmut samples
• t-test of IF vs TL samples
• Correlation of DEP TP53 analysis with RNAseq dat of TP53 mut vs nonmut

Location: server Gade Lab/Projects/PDX_TIF_proteomics_2021/TP53_analysis_Oct2021

10/18/2021

Completed analysis using paired t-tests on IF vs TL samples. Idea was to identify those proteins that had higher LFQ values in IF sample vs TL sample. Identified 98 proteins where that was the case using non-adjusted p-values. Repeated analysis using only one-tailed t-test, which yielded 141 genes.

10/6/2021

Combined gene list was created using the parameter comparison and the correlation between IF and TL using the  1_mixed_MLE_zero_1 paramater set. This could probably be improved by performing the statistical analysis on the entire dataset rather than keeping the TL and IF datasets separate, but there are issues with that given the very different distribution of NAs.

The combined list of genes is therefore composed of

1. proteins that are often statistically significantly upregulated in TP53 mut vs nonmut regardless of parameters set for the analysis. The cutoff applied was 10 parameter sets. In other words, a protein was included if it was significant in at least 10 out of 360 parameter combinations.
2. Proteins that are statistically significantly upregulated in TP53 vs nonmut samples in either TL or IF samples using non-adjusted p-values AND had log2FC of > 2 in both TL and IF comparisons.
3. 23 On/Off proteins from previous analyses

All these proteins were then graphed across sample types using both untransformed and log scales and 6 proteins were removed. For most cases analysis seems to have been biased due to outliers for those 6 proteins.

10/4/2021

After comparing parameters, gene lists were generated for enrichment analysis. However, still want to add proteins that may have not been statistically significant in either analysis, but are sufficiently correlated betweeen TL and IF that should be included also.

To do this, need results table that uses same parameters. For this 1_mixed_MLE_zero_1 was selected since these parameters give highest significant gene number for both sample types, highest sum in IF samples and second highest sum in TL samples.

10/1/2021