An increasing number of ecotoxicological studies have used omics-data to understand the dose-response patterns of environmental stressors. However, very few have investigated complex non-monotonic dose-response patterns with multi-omics data. In the present study, we developed a novel semi-supervised network analysis workflow as an alternative to benchmark dose (BMD) modelling. We utilised a previously published multi-omics dataset generated from after chronic gamma radiation exposure to obtain novel knowledge on the dose-dependent effects of radiation. Our approach combines 1) unsupervised co-expression network analysis to group genes with similar dose responses into modules; 2) supervised classification of these modules by relevant response patterns; 3) reconstruction of regulatory networks based on transcription factor binding motifs to reveal the mechanistic underpinning of the modules; 4) differential co-expression network analysis to compare the discovered modules across two datasets with different exposure periods; and 5) pathway enrichment analysis to integrate transcriptomics and metabolomics data. Our method unveiled both known and novel effects of gamma radiation, provide insight into shifts in responses from low to high dose rates, and can be used as an alternative approach for multi-omics dose-response analysis in future. The workflow SOLA (Semi-supervised Omics Landscape Analysis) is available at https://gitlab.com/wanxin.lai/SOLA.git.