Context. Infrared dark clouds (IRDCs) provide a useful testbed in which to investigate the genuine initial conditions and early stages of massive-star formation. Aims. We attempt to characterise the chemical properties of a sample of 35 massive clumps of IRDCs through multi-molecular line observations. We also search for possible evolutionary trends among the derived chemical parameters.Methods. The clumps are studied using the MALT90 (Millimetre Astronomy Legacy Team 90 GHz) line survey data obtained with the Mopra 22 m telescope. The survey covers 16 different transitions near 90 GHz. The spectral-line data are used in concert with our previous LABOCA (Large APEX BOlometer CAmera) 870 μm dust emission data. Results. Eleven MALT90 transitions are detected towards the clumps at least at the 3σ level. Most of the detected species (SiO, C 2 H, HNCO, HCN, HCO + , HNC, HC 3 N, and N 2 H + ) show spatially extended emission towards many of the sources. Most of the fractional abundances of the molecules with respect to H 2 are found to be comparable to those determined in other recent similar studies of IRDC clumps. We found that the abundances of SiO, HNCO, and HCO + are higher in IR-bright clumps than in IR-dark sources, reflecting a possible evolutionary trend. A hint of this trend is also seen for HNC and HC 3 N. An opposite trend is seen for the C 2 H and N 2 H + abundances. Moreover, a positive correlation is found between the abundances of HCO + and HNC, and between those of HNC and HCN. The HCN and HNC abundances also appear to increase as a function of the N 2 H + abundance. The HNC/HCN and N 2 H + /HNC abundance ratios are derived to be near unity on average, while that of HC 3 N/HCN is ∼10%. The N 2 H + /HNC ratio appears to increase as the clump evolves, while the HNC/HCO + ratio shows the opposite behaviour. Conclusions. The detected SiO emission is probably caused by shocks driven by outflows in most cases, although shocks resulting from the cloud formation process could also play a role. Shock-origin for the HNCO, HC 3 N, and CH 3 CN emission is also plausible. The average HNC/HCN ratio is in good agreement with those seen in other IRDCs, but gas temperature measurements would be neeeded to study its temperature dependence. Our results support the finding that C 2 H can trace the cold gas, and not just the photodissociation regions. The HC 3 N/HCN ratio appears to be comparable to the values seen in other types of objects, such as T Tauri disks and comets.