Xpress

Basic Expression Gene List Goals

Yardstick Genes

Most common problem is determining baselines and comparing experiments.

• Were these experiments measured in a different unit? TPM, RPKM, FPKM, log2 or log10 transformed?

Use the large GTEx data set to find genes that express at consistent levels across the tissue types. Some desired features:

• Common Genes - ones that are often measured.
• Consistent expression across all tissue types and samples: ** Poor information for tissue type classification.
• Avoid very low expression genes.
• Have a range of expressions levels - for internal, within sample normalization.
• Have good correlation between mRNA and protein expression levels. ** this is not always that common.

Random Point notes

1. Expression can be complicated:
   • Cite GSC UoA - decrease mRNA resulting more expression - Overall ratios of all transcripts.
2. Toxic Expression Levels - mutually exclusive genes - overactive pathway.
3. Pathway level expressions and categories?

Other effects

1. Low expression levels are irrelvant for very active proteins.
   • Activating mutations.
   • Activating fusions
   • Can pathway activation / deactivation be used to identify?
2. Micro-environment effects - cell vs tumour micro env
3. PDL1 vs mutation burden - patterns for tumour immune attack.