Limits...
Tumor evolutionary directed graphs and the history of chronic lymphocytic leukemia.

Wang J, Khiabanian H, Rossi D, Fabbri G, Gattei V, Forconi F, Laurenti L, Marasca R, Del Poeta G, Foà R, Pasqualucci L, Gaidano G, Rabadan R - Elife (2014)

Bottom Line: Cancer is a clonal evolutionary process, caused by successive accumulation of genetic alterations providing milestones of tumor initiation, progression, dissemination, and/or resistance to certain therapeutic regimes.To unravel these milestones we propose a framework, tumor evolutionary directed graphs (TEDG), which is able to characterize the history of genetic alterations by integrating longitudinal and cross-sectional genomic data.Our results suggest that TEDG may constitute an effective framework to recapitulate the evolutionary history of tumors.

View Article: PubMed Central - PubMed

Affiliation: Department of Biomedical Informatics, Columbia University, New York, United States.

ABSTRACT
Cancer is a clonal evolutionary process, caused by successive accumulation of genetic alterations providing milestones of tumor initiation, progression, dissemination, and/or resistance to certain therapeutic regimes. To unravel these milestones we propose a framework, tumor evolutionary directed graphs (TEDG), which is able to characterize the history of genetic alterations by integrating longitudinal and cross-sectional genomic data. We applied TEDG to a chronic lymphocytic leukemia (CLL) cohort of 70 patients spanning 12 years and show that: (a) the evolution of CLL follows a time-ordered process represented as a global flow in TEDG that proceeds from initiating events to late events; (b) there are two distinct and mutually exclusive evolutionary paths of CLL evolution; (c) higher fitness clones are present in later stages of the disease, indicating a progressive clonal replacement with more aggressive clones. Our results suggest that TEDG may constitute an effective framework to recapitulate the evolutionary history of tumors.

Show MeSH

Related in: MedlinePlus

Adjustment of MAF based on copy number data.(A) Definition of mutation cell frequency. The black lines within the circles represent DNA copies, and the crosses represent point mutations. The contingent table shows the difference between MAF and MCF. MAF: mutation allele frequency; MCF: mutation cell frequency; NAN: not available. (B) Optimization of Hill function by grid-search method. z-axis indicates the objective function F, x-axis and y-axis are parameters of the Hill function. (C) The optimal Hill function and the simple piecewise function. (D) MAF and MCF of the cancer two-hit model. (E) Justification of MCF. x-axis indicates the fraction of CD19+CD5+ cells assessed by FACS analysis, and y-axis indicates the maximal mutation fraction of all targeted driver genes of each sample calculated by different methods. One blue dot represents one sample, and contours indicate the density of dots. A suitable calculation of maximal driver mutation fraction will approximate but not exceed the fraction of cancer nuclei. The upper red line indicates CD19+CD5+ cell fraction, and the lower red line indicates a 20% lower interval of it. Apparently, tumor purities of 55 samples are properly assessed by the Hill function MCF, which is better than both MAF without adjustment (10 samples) and simple piecewise MCF (27 samples).DOI:http://dx.doi.org/10.7554/eLife.02869.004
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4308685&req=5

fig1s1: Adjustment of MAF based on copy number data.(A) Definition of mutation cell frequency. The black lines within the circles represent DNA copies, and the crosses represent point mutations. The contingent table shows the difference between MAF and MCF. MAF: mutation allele frequency; MCF: mutation cell frequency; NAN: not available. (B) Optimization of Hill function by grid-search method. z-axis indicates the objective function F, x-axis and y-axis are parameters of the Hill function. (C) The optimal Hill function and the simple piecewise function. (D) MAF and MCF of the cancer two-hit model. (E) Justification of MCF. x-axis indicates the fraction of CD19+CD5+ cells assessed by FACS analysis, and y-axis indicates the maximal mutation fraction of all targeted driver genes of each sample calculated by different methods. One blue dot represents one sample, and contours indicate the density of dots. A suitable calculation of maximal driver mutation fraction will approximate but not exceed the fraction of cancer nuclei. The upper red line indicates CD19+CD5+ cell fraction, and the lower red line indicates a 20% lower interval of it. Apparently, tumor purities of 55 samples are properly assessed by the Hill function MCF, which is better than both MAF without adjustment (10 samples) and simple piecewise MCF (27 samples).DOI:http://dx.doi.org/10.7554/eLife.02869.004

Mentions: To recapitulate and compare the history of genetic alterations in many patients, we propose a framework to infer TEDG by integrating longitudinal and cross-sectional genomic data of cancer patients. First, we reconstruct the sequential network of genetic alterations in each patient by analyzing genomic data from different time points. Specifically, the techniques of high-depth next generation sequencing (NGS) and fluorescence in situ hybridization (FISH) are separately carried out to assess the mutation allele frequency (MAF) and copy number abnormalities (CNA) of selected driver genes. To unify both types of data, and to adjust the MAF of mutations in genes with CNA, we introduce mutation cell frequency (MCF, defined as the fraction of tumor cells with a particular alteration) for quantification of genetic lesions (‘Materials and methods’, Figure 1—figure supplement 1). Based on MCF, we investigate alterations represented in at least 5% of leukemic cells (see examples of CLL patients in Figure 1—figure supplement 2). First, if a given genetic lesion is observed to be temporally earlier than another lesion, we connect them with a directed edge to represent their sequential order of development (Figure 1A). Second, we pool many sequential networks from different patients to construct an Integrated Sequential Network (ISN). Third, we infer TEDG from ISN by removing indirect associations with spectral techniques and minimal spanning tree algorithm. TEDG is the backbone of ISN, representing an optimal explanation of the mutation order across many patients (Figure 1B).10.7554/eLife.02869.003Figure 1.Tumor Evolutionary Directed Graph (TEDG) framework.


Tumor evolutionary directed graphs and the history of chronic lymphocytic leukemia.

Wang J, Khiabanian H, Rossi D, Fabbri G, Gattei V, Forconi F, Laurenti L, Marasca R, Del Poeta G, Foà R, Pasqualucci L, Gaidano G, Rabadan R - Elife (2014)

Adjustment of MAF based on copy number data.(A) Definition of mutation cell frequency. The black lines within the circles represent DNA copies, and the crosses represent point mutations. The contingent table shows the difference between MAF and MCF. MAF: mutation allele frequency; MCF: mutation cell frequency; NAN: not available. (B) Optimization of Hill function by grid-search method. z-axis indicates the objective function F, x-axis and y-axis are parameters of the Hill function. (C) The optimal Hill function and the simple piecewise function. (D) MAF and MCF of the cancer two-hit model. (E) Justification of MCF. x-axis indicates the fraction of CD19+CD5+ cells assessed by FACS analysis, and y-axis indicates the maximal mutation fraction of all targeted driver genes of each sample calculated by different methods. One blue dot represents one sample, and contours indicate the density of dots. A suitable calculation of maximal driver mutation fraction will approximate but not exceed the fraction of cancer nuclei. The upper red line indicates CD19+CD5+ cell fraction, and the lower red line indicates a 20% lower interval of it. Apparently, tumor purities of 55 samples are properly assessed by the Hill function MCF, which is better than both MAF without adjustment (10 samples) and simple piecewise MCF (27 samples).DOI:http://dx.doi.org/10.7554/eLife.02869.004
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4308685&req=5

fig1s1: Adjustment of MAF based on copy number data.(A) Definition of mutation cell frequency. The black lines within the circles represent DNA copies, and the crosses represent point mutations. The contingent table shows the difference between MAF and MCF. MAF: mutation allele frequency; MCF: mutation cell frequency; NAN: not available. (B) Optimization of Hill function by grid-search method. z-axis indicates the objective function F, x-axis and y-axis are parameters of the Hill function. (C) The optimal Hill function and the simple piecewise function. (D) MAF and MCF of the cancer two-hit model. (E) Justification of MCF. x-axis indicates the fraction of CD19+CD5+ cells assessed by FACS analysis, and y-axis indicates the maximal mutation fraction of all targeted driver genes of each sample calculated by different methods. One blue dot represents one sample, and contours indicate the density of dots. A suitable calculation of maximal driver mutation fraction will approximate but not exceed the fraction of cancer nuclei. The upper red line indicates CD19+CD5+ cell fraction, and the lower red line indicates a 20% lower interval of it. Apparently, tumor purities of 55 samples are properly assessed by the Hill function MCF, which is better than both MAF without adjustment (10 samples) and simple piecewise MCF (27 samples).DOI:http://dx.doi.org/10.7554/eLife.02869.004
Mentions: To recapitulate and compare the history of genetic alterations in many patients, we propose a framework to infer TEDG by integrating longitudinal and cross-sectional genomic data of cancer patients. First, we reconstruct the sequential network of genetic alterations in each patient by analyzing genomic data from different time points. Specifically, the techniques of high-depth next generation sequencing (NGS) and fluorescence in situ hybridization (FISH) are separately carried out to assess the mutation allele frequency (MAF) and copy number abnormalities (CNA) of selected driver genes. To unify both types of data, and to adjust the MAF of mutations in genes with CNA, we introduce mutation cell frequency (MCF, defined as the fraction of tumor cells with a particular alteration) for quantification of genetic lesions (‘Materials and methods’, Figure 1—figure supplement 1). Based on MCF, we investigate alterations represented in at least 5% of leukemic cells (see examples of CLL patients in Figure 1—figure supplement 2). First, if a given genetic lesion is observed to be temporally earlier than another lesion, we connect them with a directed edge to represent their sequential order of development (Figure 1A). Second, we pool many sequential networks from different patients to construct an Integrated Sequential Network (ISN). Third, we infer TEDG from ISN by removing indirect associations with spectral techniques and minimal spanning tree algorithm. TEDG is the backbone of ISN, representing an optimal explanation of the mutation order across many patients (Figure 1B).10.7554/eLife.02869.003Figure 1.Tumor Evolutionary Directed Graph (TEDG) framework.

Bottom Line: Cancer is a clonal evolutionary process, caused by successive accumulation of genetic alterations providing milestones of tumor initiation, progression, dissemination, and/or resistance to certain therapeutic regimes.To unravel these milestones we propose a framework, tumor evolutionary directed graphs (TEDG), which is able to characterize the history of genetic alterations by integrating longitudinal and cross-sectional genomic data.Our results suggest that TEDG may constitute an effective framework to recapitulate the evolutionary history of tumors.

View Article: PubMed Central - PubMed

Affiliation: Department of Biomedical Informatics, Columbia University, New York, United States.

ABSTRACT
Cancer is a clonal evolutionary process, caused by successive accumulation of genetic alterations providing milestones of tumor initiation, progression, dissemination, and/or resistance to certain therapeutic regimes. To unravel these milestones we propose a framework, tumor evolutionary directed graphs (TEDG), which is able to characterize the history of genetic alterations by integrating longitudinal and cross-sectional genomic data. We applied TEDG to a chronic lymphocytic leukemia (CLL) cohort of 70 patients spanning 12 years and show that: (a) the evolution of CLL follows a time-ordered process represented as a global flow in TEDG that proceeds from initiating events to late events; (b) there are two distinct and mutually exclusive evolutionary paths of CLL evolution; (c) higher fitness clones are present in later stages of the disease, indicating a progressive clonal replacement with more aggressive clones. Our results suggest that TEDG may constitute an effective framework to recapitulate the evolutionary history of tumors.

Show MeSH
Related in: MedlinePlus