Deep profiling of antibody and T cell-receptor repertoires by means of high-throughput sequencing has become an attractive approach for adaptive immunity studies, but its power is substantially compromised by the accumulation of PCR and sequencing errors. Here we report MIGEC (molecular identifier groups-based error correction), a strategy for high-throughput sequencing data analysis. MIGEC allows for nearly absolute error correction while fully preserving the natural diversity of complex immune repertoires.