Biology subjects

Leiserson, M.

Publications and source records attributed to Leiserson, M..

3 recordsLinked to original sources

bioRxiv2018

Hidden Markov Models Lead to Higher Resolution Maps of Mutation Signature Activity in Cancer

Knowing the activity of the mutational processes shaping a cancer genome may provide insight into tumorigenesis and personalized therapy. It is thus important to uncover the characteristic signatures of active mutational processes in patients from their patterns of single base substitutions. However, mutational processes do not act uniformly on the genome and are biased by factors such as the genomes chromatin structure or replication origins. These factors may lead to statistical dependencies among neighboring mutations, calling for modeling approaches that can account for such dependencies to better estimate mutational process activities.\n\nHere we develop the first sequence-dependent models for mutation signatures. We apply these models to characterize genomic and other factors that influence the activity of previously validated mutation signatures in breast cancer. We find that our tool, SO_SCPLOWIGC_SCPLOWMO_SCPLOWAC_SCPLOW, can accurately assign genomic mutations to mutation signatures, yielding assignments that are of higher likelihood than those obtained with models that assume independence between signatures and align better with current biological knowledge. Our analysis resolves a controversy related to the dependency of APOBEC signatures on replication time and links Signatures 18 and 30 to oxidative damage.\n\nModeling the sequential dependencies of mutation signatures leads to improved estimates of mutation signature activity both at the tumor-level and within specific genomic regions, yielding higher resolution maps of mutation signature activity in cancer.

bioinformatics↗

bioRxiv2018

Criticality in Tumor Evolution and Clinical Outcome

How mutation and selection determine the fitness landscape of tumors and hence clinical outcome is an open fundamental question in cancer biology, crucial for the assessment of therapeutic strategies and resistance to treatment. Here we explore the mutation-selection phase-diagram of 6721 primary tumors representing 23 cancer types, by quantifying the overall somatic point mutation load (ML) and selection (dN/dS) in the entire proteome of each tumor. We show that ML strongly correlates with patient survival, revealing two opposing regimes around a critical point. In low ML cancers, high number of mutations indicates poor prognosis, whereas high ML cancers show the opposite trend, due to mutational meltdown. Although the majority of cancers evolve near neutrality, deviations are observed at extreme MLs. Cancers with the highest ML evolve under purifying selection, whereas those with the lowest ML show signatures of positive selection, demonstrating how selection affects cancer fitness. Moreover, different cancers occupy specific positions on the ML-dN/dS plane, revealing a diversity of evolutionary trajectories. These results support and expand the theory of tumor evolution and its non-linear effects on survival.\n\nSignificance StatementIt remains an open fundamental question how mutation and selection co-determine the course of cancer evolution. We construct a selection-mutation phase diagram, using tumor mutation load and selection strength as key variables, and assess their association with clinical outcome. We demonstrate the existence of a biphasic evolutionary regime, whereby beyond a critical ML, the fitness of tumors decreases with the number of mutations, while the proteome evolves near neutrality. Deviations from neutrality in extreme ML elucidate how positive and purifying selections maintain tumor fitness. These results empirically corroborate the existence of a critical state in cancer evolution predicted by theory, and have fundamental and likely clinical implications.

cancer biology↗

bioRxiv2017

A Multi-Species Functional Embedding Integrating Sequence and Network Structure

A key challenge to transferring knowledge between species is that different species have fundamentally different genetic architectures. Initial computational approaches to transfer knowledge across species have relied on measures of heredity such as genetic homology, but these approaches suffer from limitations. First, only a small subset of genes have homologs, limiting the amount of knowledge that can be transferred, and second, genes change or repurpose functions, complicating the transfer of knowledge. Many approaches address this problem by expanding the notion of homology by leveraging high-throughput genomic and proteomic measurements, such as through network alignment.\n\nIn this work, we take a new approach to transferring knowledge across species by expanding the notion of homology through explicit measures of functional similarity between proteins in different species. Specifically, our kernel-based method, HO_SCPLOWANDLC_SCPLOW (Homology Assessment across Networks using Diffusion and Landmarks), integrates sequence and network structure to create a functional embedding in which proteins from different species are embedded in the same vector space. We show that inner products in this space capture functional similarity across species, and the vectors themselves are useful for a variety of cross species tasks. We perform the first whole-genome method for predicting phenologs, generating many that were previously identified, but also predicting new phenologs supported from the biological literature. We also demonstrate the HO_SCPLOWANDLC_SCPLOW-embedding captures pairwise gene function, in that gene pairs with synthetic lethal interactions are co-located in HO_SCPLOWANDLC_SCPLOW-space both within and across species. Software for the HO_SCPLOWANDLC_SCPLOW algorithm is available at http://github.com/lrgr/HANDL.

bioinformatics↗