Central to the different forms of taxis are methyl-accepting chemotaxis proteins (MCPs). The increasing number of genome sequences reveals that MCPs differ enormously in sequence, topology and genomic abundance. This work is a one-by-one bioinformatic analysis of the almost-totality of MCP genes available and a classification of motile bacteria according to their lifestyle. On average, motile archaea have 6.7 MCP genes per genome whereas motile bacteria have more than twice as much. We show that the number of MCPs per genome depends on bacterial lifestyle and metabolic diversity, but weakly on genome size. Signal perception at an MCP occurs at the N-terminal ligand binding region (LBR). Here we show that around 88% of MCPs possess an LBR that remains un-annotated in SMART. MCPs can be classified into two clusters according to the size of the LBR. Cluster I receptors have an LBR between 120 and 210 amino acids whereas cluster II receptors have larger LBRs of 220–299 amino acids. There is evidence that suggests that some cluster II LBRs are composed of two cluster I LBRs. Further evidence indicates that other cluster II LBRs might harbour novel sensor domains. Cluster II receptors are dominant in archaea whereas cluster I receptors are prevalent in bacteria. MCPs can be classified into six different receptor topologies and this work contains a first estimation of the relative abundance of different receptor topologies in bacteria and archaea. Topologies involving extracytoplasmic sensing are prevalent in bacteria whereas topologies with cytosolic signal recognition are abundant in archaea.