DDRi_RNAseq

RNAseq analysis repository for analysis

Experimental Design

Samples and Conditions

Each sample has two key attributes:

• Gene Target (e.g. GAL4, PRDX1, NTC, etc.)
• DNAPKi Treatment (TRUE/FALSE)

We combine these into a single factor called condition. For example:

• GAL4_FALSE = "GAL4 with no DNAPKi treatment"
• GAL4_TRUE = "GAL4 with DNAPKi treatment"
• PRDX1_FALSE = "PRDX1 with no DNAPKi treatment"
• PRDX1_TRUE = "PRDX1 with DNAPKi treatment"
• NTC_FALSE = "Negative control (no DNAPKi)"

Reference Level

We make NTC_FALSE the reference (baseline) level. This ensures that any coefficient for "conditionXYZ" in the linear model is interpreted as:

(XYZ) - (NTC_FALSE)

Design Matrix

We set up a linear model with the formula ~ replicate + condition, where:

• replicate accounts for batch or biological replicate effects
• condition captures all combinations of gene target and DNAPKi status

Concretely, our design matrix columns include:

• Intercept (which corresponds to NTC_FALSE)
• Replicate terms
• conditionGAL4_FALSE, conditionGAL4_TRUE, conditionPRDX1_FALSE, conditionPRDX1_TRUE, etc.

Single Comparisons

For any given conditionX, we do a single comparison X vs NTC_FALSE by extracting that specific coefficient in the model. This tells us whether X is significantly different from the negative control (no DNAPKi).

Double Comparisons (DNAPKi Effects)

We define contrasts to compare (GeneTarget_TRUE - GeneTarget_FALSE).

For example:

• GAL4_DNAPKi_effect = conditionGAL4_TRUE - conditionGAL4_FALSE
• PRDX1_DNAPKi_effect = conditionPRDX1_TRUE - conditionPRDX1_FALSE

Each such contrast tests whether DNAPKi treatment (for a specific gene target) significantly changes expression compared to the same gene target without DNAPKi.

Interpretation

Single Comparison vs. NTC_FALSE:

• logFC > 0 ⇒ Genes are up-regulated relative to the negative control
• logFC < 0 ⇒ Genes are down-regulated relative to the negative control

Double Comparison (DNAPKi effect):

• logFC > 0 ⇒ Genes are further increased by DNAPKi treatment (for that gene target)
• logFC < 0 ⇒ Genes are decreased by DNAPKi for that gene target