FPKM

FPKM, or fragments per kilobase of transcript per million mapped fragments, is a normalization metric used in RNA-seq to estimate gene expression levels. It accounts for both sequencing depth and gene length, enabling comparisons of expression within a sample between genes of different lengths.

Calculation: Let C be the number of fragments mapping to a gene, N the total number of mapped fragments in the sample, and L the transcript length in base pairs. The FPKM for the gene is FPKM = (C × 1,000,000,000) / (N × L). Equivalently, using L in kilobases, FPKM = (C / L_k) / (N / 1,000,000), where L_k is the gene length in kilobases.

Uses: FPKM provides a normalized estimate of expression suitable for ranking genes by abundance within a sample. It is commonly reported in RNA-seq pipelines and public resources that use transcript-level quantification.

Limitations: FPKM is sensitive to library composition and is not ideal for comparing expression across samples for differential expression analyses, due to compositional biases and varying isoform structures. TPM (transcripts per million) and robust count-based methods (e.g., DESeq2, edgeR) are generally preferred for cross-sample comparisons and differential expression. Additionally, annotation changes and multi-mapped reads can affect FPKM values.

History: The concept derives from RPKM/FPKM introduced in early RNA-seq studies; many modern analyses favor TPM or count-based approaches for statistical testing.