Permafrost environments currently sequester two times more carbon than the atmosphere. However, climate change is leading to permafrost thaw which allows organic carbon to be degraded by microorganisms and the release of greenhouse gasses such as carbon dioxide (CO2) and methane (CH4). Despite the global importance of this process, organic carbon degradation and the microorganisms responsible are poorly understood. Over the last 30 years, 10% of intact permafrost at Stordalen Mire, Sweden, has thawed leading to the formation of greenhouse gas emitting wetland environments. This site provides an opportunity to understand organic carbon breakdown in the context of a natural permafrost thaw gradient. Here, metagenomic and metatranscriptomic sequencing of microbial communities was coupled to soil organic matter (SOM) characterisation at key stages of permafrost thaw at Stordalen Mire. SOM was characterised at the molecular level using electrospray ionization (ESI) coupled with high resolution Fourier Transform ion cyclotron resonance mass spectrometry (FTICR MS), providing insight into the complex molecular composition of metabolites along the thaw gradient. This approach allows the metagenome, transcriptome and SOM-based measurements to be combined into a single transformation network by overlaying the relative abundance and expression of genes in a metagenome and transcriptome whose products mediate specific SOM transformations from the same sample. Using this approach, it was found that partially thawed environments in Stordalen Mire had significantly higher abundances of sugars such as glucose, mannose, and xylose, and lower abundances of enzymes associated with their uptake and metabolism. Fully thawed environments were characterised by lower abundances of sugars, complex metabolism and a higher net release of the greenhouse gas CH4. This metabolic network construction based on unbiased SOM and -omic composition offers a breakthrough in systems-wide analysis of microbial metabolisms in soil.