Introduction: Small Cell Lung Cancer (SCLC) is characterized by aggressive behavior
and limited
treatment efficacy. The integration of proteomics and gene expression data offers
new avenues for predicting drug
responses. This study utilizes dimension reduction and clustering techniques to identify
distinct sample clusters
within SCLC cell lines, aiming to discover variable drug responses and enhance personalized
treatment strategies.
Methods and Results: Proteomic and gene expression data from SCLC cell lines were
subjected to
dimension reduction techniques (tSNE, UMAP, PCA) and clustering methods (DynamicTreeCut,
HDBSCAN, k-means)
to form distinct sample clusters. Drug response data, in terms of IC50 values, were
obtained from databases such
as NCI, GDSC, CCLE, and UTSW. Within each sample cluster, drug responses were analyzed
to identify patterns of
sensitivity or resistance. Additionally, ssGSEA was employed to derive enrichment
scores from gene clusters, which
were then correlated with the drug response data to pinpoint biomarkers predictive
of treatment outcomes. This
dual approach of clustering samples and correlating gene cluster enrichment scores
with drug responses enabled
the identification of distinct phenotypic profiles with specific drug susceptibilities.
Conclusion: By clustering SCLC cell lines based on proteomic and transcriptomic data
and correlating
gene expression signatures with drug responses, our study delineates a comprehensive
method to discern and
exploit therapeutic vulnerabilities in SCLC. This strategy enhances the precision
of drug response p
