CALDER: Inferring Phylogenetic Trees from Longitudinal Tumor Samples

Abstract

Highlights d Longitudinal sequencing provides additional information for phylogeny inference d CALDER leverages longitudinal information to derive phylogeny from mixed samples d CALDER yields more accurate trees on simulated and real cancer data d Longitudinal model extendable to other data types such as single-cell sequencing

Longitudinal cancer evolution from single cells

“…We generated a total of 450 independent datasets for distinct experimental scenarios (see below) and compared LACE with SCITE [13] and TRaIT [16], two high-performing tools for the inference of mutational trees from single-cell sequencing data on single time points, and with CALDER [19], the benchmark tool for longitudinal phylogenetic tree inference from bulk sequencing data. As synthetic datasets need to be adapted to be processed by such tools, with respect to CALDER, we computed the cancer cell fraction of driver mutations from the observed single-cell genotypes, and by assuming a uniform sampling of single cells and a read depth of 200X, which is typical for whole-exome sequencing experiments.…”

“…On the one hand, in fact, the existing list of approaches that process single-cell data and extend phylogenetic methods by handling data-specific errors [13,14,15,16], are not suitable to handle multiple temporally ordered samples derived from the same tumor, and cannot be used to investigate the clonal prevalence variation in time. On the other hand, even though methods for longitudinal bulk sequencing data are starting to produce noteworthy results [17,18,19], they usually require complex computational strategies to deconvolve the signal coming from intermixed cell subpopulations. Furthermore, there is an ongoing debate whether multi-sample trees from bulk samples are indeed phylogenies or, conversely, if they might lead to erroneous evolutionary inferences [20].…”

“…In order to assess the accuracy and robustness of the results produced by LACE, we performed extensive simulations, and compared with CALDER [19], a recent method for the reconstruction of longitudinal phylogenetic trees from bulk samples, SCITE [13] and TRaIT [16], two state-of-the-art tools for the inference of mutational trees from single-cell sequencing data.…”

Single-cell tumor phylogeny inference with copy-number constrained mutation losses

“…We evaluated the phylogenies output by the methods by two measures that have been previously used in tumor evolution studies 11,15,23,28,29,41 . First, the mutation matrix error M (B,B) = 1 mn m i=1 n j=1 |b i,j − b i,j | is the normalized Hamming distance between the inferred binary mutation matrixB and the true binary mutation matrix B and assesses the accuracy of the error-corrected mutation profiles for each observed cell.…”

“…The evolution of a tumor is typically described by a phylogenetic tree, or phylogeny, whose leaves represent the cells observed at the present time and whose internal nodes represent ancestral cells. Tumor phylogenies are challenging to reconstruct using DNA sequencing data from bulk tumor samples, since this data contains mixtures of mutations from thousandsmillions of heterogeneous cells in the sample [3][4][5][6][7][8][9][10][11][12][13][14][15] . Recently, single-cell DNA sequencing (scDNA-seq) of tumors has become more common, and new technologies such as those from 10X Genomics 16…”