Leveraging sequences missing from the human genome to diagnose cancer

Abstract

Cancer diagnosis using cell-free DNA (cfDNA) can significantly improve treatment and survival but has several technical limitations. Here, we show that tumor-associated mutations create neomers, DNA sequences 11-18bp in length that are absent in the human genome, that can accurately detect cancer subtypes and features. We show that we can detect twenty-one different tumor-types with higher accuracy than state-of-the-art methods using a neomer-based classifier. Refinement of this classifier via supervised learning identified additional cancer features with even greater precision. We also demonstrate that neomers can precisely diagnose cancer from cfDNA in liquid biopsy samples. Finally, we show that neomers can be used to detect cancer-associated non-coding mutations affecting gene regulatory activity. Combined, our results identify a novel, sensitive, specific and straightforward cancer diagnostic tool.

Frequentmers - a novel way to look at metagenomic Next Generation Sequencing data and an application in detecting liver cirrhosis

“…We extracted every sixteen base-pair (bp) kmer found in each sample and split samples in ten groups or folds, with the proportion of cases over the total samples in each fold being maintained. The choice of kmer length was informed from previous studies in which we found that the performance of the kmer-based models increased as a function of kmer length up to sixteen bp length (Georgakopoulos-Soares, Barnea, et al 2021). For each fold, we examined which subset of the total kmers detected constituted frequentmers, using the number of samples in which each kmer was found as the recurrency threshold (see Methods).…”

“…Kmers have been previously used to describe new features or characteristics of an organism related to the presence or absence of a specific contiguous subsequence. For example, the subset of kmers that do not appear in a genome are referred to as nullomers (Acquisti et al 2007; Georgakopoulos-Soares, Yizhar-Barnea, et al 2021; Koulouras and Frith 2021) and the subset of kmers that are found in a single species are referred to as quasi-primes (Mouratidis et al 2023). Using kmer strategies, we may efficiently mine the human genome for differences that distinguish patients with disease from healthy individuals in effort to establish unique biological signatures.…”

kmerDB: A Database Encompassing the Set of Genomic and Proteomic Sequence Information for Each Species

“…Identification of kmers was performed as previously defined in (Georgakopoulos-Soares, Yizhar-Barnea, et al 2021). For nucleic kmers and nullomers, the lengths of six to twelve bps were used for eukaryotes, bacteria, and archaea, whereas for viruses due to their smaller genome size, the lengths of three to six bps were used.…”

“…For peptide kmers, oligopeptide lengths of up to seven amino acids were used across all the species. Nullomer and nullpeptide detection was performed as previously described in (Georgakopoulos-Soares, Yizhar-Barnea, et al 2021) for each species at each kmer length.…”

“…Genomes and proteomes consist of sequences of oligonucleotides and oligopeptides, respectively, which can be partitioned into substrings of a fixed length k, known as kmers. Among these kmers, some are conspicuously absent from a given genome or proteome, and are termed nullomers or nullpeptides (Acquisti et al 2007; Herold, Kurtz, and Giegerich 2008; Koulouras and Frith 2021; Georgakopoulos-Soares, Yizhar-Barnea, et al 2021) These kmer sequences have been used for various applications, such as quality control, metagenomics classification, phylogenetic analysis, binning, evolution studies, and identification of pathogens (Jagadeesan et al 2019; Deurenberg et al 2017; Brandies et al 2019; Maljkovic Berry et al 2020). Remarkably, the resurfacing of nullomers in the human genome has been leveraged to detect cancer (Georgakopoulos-Soares, Barnea, et al 2021; Montgomery et al 2023), demonstrating their potential for cancer and disease diagnostics.…”