I have been working on getting the differential expression of proteins for 2 groups - control and treatment. Control has 4 animals while treatment has 5 animals. The samples are collected at the baseline time and after 2 hours. The serum was used to do a somascan assay. I have been using the limma model to get the differential expression of proteins but I am constantly getting an error after I apply the limma function. It says this - Error in lmFit(protein_data_log2, design) : row dimension of design doesn't match column dimension of data object after I use this prompt - fit <- lmFit(protein_data_log2, design)

Code should be placed in three backticks as shown below

library(limma)
library(dplyr)

soma_data <- read.csv("20241220.csv")

print(paste("Number of rows in soma_data:", nrow(soma_data)))
print(paste("Number of columns in soma_data:", ncol(soma_data)))
print(head(soma_data))

unique_ids <- make.unique(as.character(soma_data$SampleId))
print(paste("Number of unique_ids:", length(unique_ids)))

protein_cols <- grep("seq\\.", colnames(soma_data))
protein_data <- soma_data[, protein_cols]
print(paste("Dimensions of protein_data (before transpose):", nrow(protein_data), "x", ncol(protein_data)))
rownames(protein_data) <- unique_ids  # Use unique_ids here!

protein_data_transposed <- t(protein_data)
print(paste("Dimensions of protein_data_transposed:", nrow(protein_data_transposed), "x", ncol(protein_data_transposed)))

protein_data_log2 <- log2(protein_data_transposed)
print(paste("Dimensions of protein_data_log2:", nrow(protein_data_log2), "x", ncol(protein_data_log2)))

sample_data <- data.frame(unique_id = unique_ids, SampleGroup = soma_data$SampleGroup)
print(paste("Dimensions of sample_data:", nrow(sample_data), "x", ncol(sample_data)))
print(head(sample_data))

ids_after_transpose = colnames(protein_data_log2)
sample_data = sample_data[match(ids_after_transpose, sample_data$unique_id),]

print(paste("Dimensions of sample_data:", nrow(sample_data), "x", ncol(sample_data)))
print(paste("Unique length of data is:", length(ids_after_transpose)))

design <- model.matrix(~0 + factor(sample_data$SampleGroup))
colnames(design) <- levels(factor(sample_data$SampleGroup))
rownames(design) <- ids_after_transpose # Align matrix!

print(paste("Dimensions of protein_data_log2 (rows x cols):", nrow(protein_data_log2), ncol(protein_data_log2)))
print(paste("Dimensions of design (rows x cols):", nrow(design), ncol(design)))
print(paste("Are row names of design and column names of protein_data_log2 equal?", all(rownames(design) == colnames(protein_data_log2))))

print(paste("Dimensions of protein_data_log2:", nrow(protein_data_log2), "rows x", ncol(protein_data_log2), "columns (samples)"))
print(paste("Dimensions of design:", nrow(design), "rows x", ncol(design), "columns (groups)"))
print(paste("Number of unique_ids:", length(unique_ids)))
print(paste("Unique length of data is:", length(ids_after_transpose)))

fit <- lmFit(protein_data_log2, design)


contrast_matrix <- makeContrasts(
  GroupAvsGroupB = "Control" - "Disease", 
  levels = design
)

fit_contrast <- contrasts.fit(fit, contrast_matrix)
fit_bayes <- eBayes(fit_contrast)

results <- topTable(fit_bayes, coef = "GroupAvsGroupB", n = Inf)
head(results)

Providing a bunch of code but not the output is not very helpful. Ideally you would just provide

nrow(design)
ncol(protein_data_log2)

And be done with it. And since lmFit says those two numbers don't match, then for sure they don't match. This also implies that the number of rows in sample_data differ from the number of columns in 'protein_data_log2`, which also means those data don't line up correctly. Ensuring that the phenotypic data and the observations are aligned is a critical preliminary step in any analysis, and you should take the time to check that.