Normalized Connectivity Score (NCS)

Summary

The Normalized Connectivity Score (NCS) is a standardized statistical metric developed by the Connectivity Map (CMap) and L1000 platforms to measure the similarity between a query gene expression signature and a reference perturbagen signature. The NCS is designed to compare up-regulated and down-regulated query genes separately against the reference, and then normalize the raw connectivity scores across different cell lines to account for variations in cell line sensitivity. In benchmarking studies, the NCS has been compared to Spearman correlation for ranking drug candidates in TWAS signature-matching.

Mathematical Formulation

The calculation of the Normalized Connectivity Score proceeds through the calculation of a Weighted Connectivity Score (WTCS) based on a weighted Kolmogorov-Smirnov-like statistic:

1. Weighted Connectivity Score (WTCS)

Given a query $q$ partitioned into up-regulated genes ($q_{up}$) and down-regulated genes ($q_{down}$), and a drug profile $d$, enrichment scores $ES_{up}$ and $ES_{down}$ are calculated. The WTCS is defined as: $$w_{q,d} = \begin{cases} \frac{ES_{up} - ES_{down}}{2} & \text{if } \text{sign}(ES_{up}) \neq \text{sign}(ES_{down}) \ 0 & \text{otherwise} \end{cases}$$

2. Normalization (NCS)

To enable comparison across different cell lines $c$, the WTCS vector $w$ is normalized by the signed means of the scores: $$NCS_c = \begin{cases} \frac{w_c}{\mu_c^+} & \text{if } \text{sign}(w_{c,t} > 0) \ \frac{w_c}{\mu_c^-} & \text{otherwise} \end{cases}$$ where $\mu_c^+$ and $\mu_c^-$ are the signed means of the raw weighted connectivity scores evaluated separately for positive and negative scores within cell line $c$. The resulting NCS ranges from -1 to 1, where a score of 1 indicates the strongest positive correlation and -1 represents the strongest negative correlation.

Benchmark Evaluations

A benchmarking study by the Shah Lab (as of 2026-07-19) demonstrated that the NCS performs poorly compared to Spearman correlation in prioritizing known first-line treatments like HMGCR inhibitors. This difference is likely because the NCS calculates enrichment scores for up- and down-regulated gene sets separately, whereas Spearman correlation evaluates the entire signature together, making Spearman correlation more robust for TWAS signature-matching.

Citations

• Lamb, J., Crawford, E. D., Peck, D., Modell, J. W., Blat, I. C., Wrobel, M. J., ... & Golub, T. R. (2006). The Connectivity Map: using gene-expression signatures to connect small molecules, genes, and disease. Science, 313(5795), 1929–1935.
• Subramanian, A., Narayan, R., Corsello, S. M., Peck, D. D., Natoli, T. E., Lu, X., ... & Golub, T. R. (2017). A next generation connectivity map: L1000 platform and the first 1,000,000 profiles. Cell, 171(6), 1437–1452.
• Chauquet, S., Jiang, J. C., Barker, L. F., Hunter, Z. L., Singh, G., Wray, N. R., McRae, A. F., & Shah, S. (2026). From GWAS to drug: A framework for drug candidate prioritisation using a gene expression signature matching approach. medRxiv. Source paper: 2026.04.22.26349470v2.full.pdf