Methods and compositions for reducing base errors of massive parallel sequencing using triseq sequencing

Methods and compositions for making DNA libraries for massive parallel next generation sequencing (NGS), comprises two parts. These methods may be referred to as Triseq sequencing. The first part includes ligating a UMI adapter, amplifying the DNA fragments in the presence of dUTP, enriching the target molecules through primer extension by using a panel of both forward and reverse primers, and removing the dU-containing template DNA. The DNA molecules are organized to primary clones and subclones, labeled by the UMI on 5′ and 3′ end of the DNA fragments, respectively. The second part includes sequencing the DNA library by NGS, deducing consensus sequence from each subclone, and from within each primary clone, and between the consensus sequences obtained from both forward and reverse primers.