BioMed Research International / 2018 / Article / Fig 6

Research Article

PATRI, a Genomics Data Integration Tool for Biomarker Discovery

Figure 6

Simulation of Clinical” analysis workflow based on presumed ALK inhibitor sensitivity in lymphoma clinical samples. Results of PATRI “Clinical” analysis workflow, simulated using two panels of lymphoma clinical samples. The 5 ALK-positive ALCL samples in the GSE14879 dataset (sample IDs: GSM368499, GSM368506, GSM368519, GSM368565, and GSM368566) were presumed to be ALK inhibitor responders and set as “Sensitive” samples only for validation purposes, using the PATRI available gene expression analysis algorithms. (a) Heatmap representing hierarchical cluster analysis via random forest categorization of the predicted “Sensitive” or “Resistant” 130 samples in the lymphoma GSE19069 dataset (sample IDs could not be represented on the lower part of the graph), starting from a filtered 17-gene expression biomarker list obtained by Limma analysis of the GSE14879 dataset (p value<10−10, logFC>|1|). (b) Histogram representing lymphoma diagnosis distribution for the 14 predicted “Sensitive” and the 116 “Resistant” lymphoma samples from the GSE19069 dataset (ALCL, ALK+: anaplastic large cell lymphoma ALK-positive; ALCL, ALK-: anaplastic large cell lymphoma ALK-negative; PTCL-NOS: peripheral T-cell lymphoma, unspecified; ATLL: adult T-cell leukemia/lymphoma; Angioimmunoblastic: angioimmunoblastic T-cell lymphoma).
(a)
(b)