A unified ILP framework for core ancestral genome reconstruction problems

Improving the identification of cancer driver modules using deep features learned from multi-omics data

Identifying the cancer driver modules or pathways is crucial to understanding the fundamental mechanisms of cancer occurrence and progression. The rapid abundance of cancer omics data provides unprecedented opportunities to study the driver modules in cancer, and many computational methods have been developed in recent years. However, most existing methods have limitations in considering different types of cancer omics data and cannot effectively learn informative omics features for integrated identification of driver modules. In this paper, we introduce a new integrated framework to accurately identify the cancer driver modules by integrating the protein-protein interaction network, transcriptional regulatory network, gene expression and mutation data in cancer. We first develop a series of methods to learn the deep features of functional connectivity between genes in each omics data and then construct an integrated gene functional coherence network. Furthermore, we present a two-step module mining method to efficiently identify the cancer driver modules from the integrated gene functional coherence network. Systematic experiments in three cancer types demonstrate that the proposed framework can obtain more significant driver modules than most existing methods, and some identified driver modules are associated with clinical survival phenotypes.

TruEst: a better estimator of evolutionary distance under the INFER model

Genome rearrangements are evolutionary events that shuffle genomic architectures. The number of genome rearrangements that happened between two genomes is often used as the evolutionary distance between these species. This number is often estimated as the minimum number of genome rearrangements required to transform one genome into another which are only reliable for closely-related genomes. These estimations often underestimate the evolutionary distance for genomes that have substantially evolved from each other, and advanced statistical methods can be used to improve accuracy. Several statistical estimators have been developed, under various evolutionary models, of which the most complete one, INFER, takes into account different degrees of genome fragility. We present TruEst-an efficient tool that estimates the evolutionary distance between the genomes under the INFER model of genome rearrangements. We apply our method to both simulated and real data. It shows high accuracy on the simulated data. On the real datasets of mammal genomes the method found several pairs of genomes for which the estimated distances are in high consistency with the previous ancestral reconstruction studies.

IAGS: Inferring Ancestor Genome Structure under a wide range of evolutionary scenarios

Significant improvements in genome sequencing and assembly technology have led to increasing numbers of high-quality genomes, revealing complex evolutionary scenarios such as multiple whole-genome duplication events, which hinders ancestral genome reconstruction via the currently available computational frameworks. Here, we present the Inferring Ancestor Genome Structure (IAGS) framework, a novel block/endpoint matching optimization strategy with single-cut-or-join distance, to allow ancestral genome reconstruction under both simple (single-copy ancestor) and complex (multicopy ancestor) scenarios. We evaluated IAGS with two simulated data sets and applied it to four different real evolutionary scenarios to demonstrate its performance and general applicability. IAGS is available at https://github.com/xjtu-omics/IAGS.

Ancestral Flowering Plant Chromosomes and Gene Orders Based on Generalized Adjacencies and Chromosomal Gene Co-Occurrences

Recurrent whole genome duplication and the ensuing loss of redundant genes-fractionation-complicate efforts to reconstruct the gene orders and chromosomes of the ancestors associated with the nodes of a phylogeny. Loss of genes disrupts the gene adjacencies key to current techniques. With our RACCROCHE pipeline, instead of starting with the inference of short ancestral segments, we suggest delaying the choice of gene adjacencies while we accumulate many more syntenically validated generalized (gapped) adjacencies. We obtain longer ancestral contigs using maximum weight matching (MWM). Similarly, we do not construct chromosomes by successively piecing together contigs into larger segments, but rather compile counts of pairwise contig co-occurrences on the set of extant genomes and use these to cluster the contigs. Chromosome-level contig assemblies for a monoploid genome emerge naturally at each node of the phylogeny and the contigs then can be ordered along the chromosome. Sampling alternative MWM solutions, visualizing heat maps, and applying gap statistics allow us to estimate the number of chromosomes in the reconstruction. We introduce several measures of quality: length of contigs, continuity of contig structure on successive ancestors, coverage of the extant genome by the reconstruction, and rearrangement relations among the inferred chromosomes. The reconstructed ancestors are visualized by painting the ancestral projections on the descendant genomes. We submit genomes drawn from a broad range of monocot orders to our pipeline, confirming the tetraploidization event "tau" in the stem lineage between the alismatids and the lilioids. We show additional applications to the Solanaceae and to four Brassica genomes, producing evidence about the monoploid ancestor in each case.

Phylostratic Shift of Whole-Genome Duplications in Normal Mammalian Tissues towards Unicellularity Is Driven by Developmental Bivalent Genes and Reveals a Link to Cancer

Tumours were recently revealed to undergo a phylostratic and phenotypic shift to unicellularity. As well, aggressive tumours are characterized by an increased proportion of polyploid cells. In order to investigate a possible shared causation of these two features, we performed a comparative phylostratigraphic analysis of ploidy-related genes, obtained from transcriptomic data for polyploid and diploid human and mouse tissues using pairwise cross-species transcriptome comparison and principal component analysis. Our results indicate that polyploidy shifts the evolutionary age balance of the expressed genes from the late metazoan phylostrata towards the upregulation of unicellular and early metazoan phylostrata. The up-regulation of unicellular metabolic and drug-resistance pathways and the downregulation of pathways related to circadian clock were identified. This evolutionary shift was associated with the enrichment of ploidy with bivalent genes (p < 10⁻¹⁶). The protein interactome of activated bivalent genes revealed the increase of the connectivity of unicellulars and (early) multicellulars, while circadian regulators were depressed. The mutual polyploidy-c-MYC-bivalent genes-associated protein network was organized by gene-hubs engaged in both embryonic development and metastatic cancer including driver (proto)-oncogenes of viral origin. Our data suggest that, in cancer, the atavistic shift goes hand-in-hand with polyploidy and is driven by epigenetic mechanisms impinging on development-related bivalent genes.