An evolution-based machine learning to identify cancer type-specific driver mutations.

Identifying cancer type-specific driver mutations is crucial for illuminating distinct pathologic mechanisms across various tumors and providing opportunities of patient-specific treatment. However, although many computational methods were developed to predict driver mutations in a type-specific manner, the methods still have room to improve. Here, we devise a novel feature based on sequence co-evolution analysis to identify cancer type-specific driver mutations and construct a machine learning (ML) model with state-of-the-art performance. Specifically, relying on 28 000 tumor samples across 66 cancer types, our ML framework outperformed current leading methods of detecting cancer driver mutations. Interestingly, the cancer mutations identified by sequence co-evolution feature are frequently observed in interfaces mediating tissue-specific protein-protein interactions that are known to associate with shaping tissue-specific oncogenesis. Moreover, we provide pre-calculated potential oncogenicity on available human proteins with prediction scores of all possible residue alterations through user-friendly website (http://sbi.postech.ac.kr/w/cancerCE). This work will facilitate the identification of cancer type-specific driver mutations in newly sequenced tumor samples.

Transcriptomes of the tumor-adjacent normal tissues are more informative than tumors in predicting recurrence in colorectal cancer patients.

BACKGROUND: Previous investigations of transcriptomic signatures of cancer patient survival and post-therapy relapse have focused on tumor tissue. In contrast, here we show that in colorectal cancer (CRC) transcriptomes derived from normal tissues adjacent to tumors (NATs) are better predictors of relapse. RESULTS: Using the transcriptomes of paired tumor and NAT specimens from 80 Korean CRC patients retrospectively determined to be in recurrence or nonrecurrence states, we found that, when comparing recurrent with nonrecurrent samples, NATs exhibit a greater number of differentially expressed genes (DEGs) than tumors. Training two prognostic elastic net-based machine learning models-NAT-based and tumor-based in our Samsung Medical Center (SMC) cohort, we found that NAT-based model performed better in predicting the survival when the model was applied to the tumor-derived transcriptomes of an independent cohort of 450 COAD patients in TCGA. Furthermore, compositions of tumor-infiltrating immune cells in NATs were found to have better prognostic capability than in tumors. We also confirmed through Cox regression analysis that in both SMC-CRC as well as in TCGA-COAD cohorts, a greater proportion of genes exhibited significant hazard ratio when NAT-derived transcriptome was used compared to when tumor-derived transcriptome was used. CONCLUSIONS: Taken together, our results strongly suggest that NAT-derived transcriptomes and immune cell composition of CRC are better predictors of patient survival and tumor recurrence than the primary tumor.

Divergence of Noncoding Regulatory Elements Explains Gene-Phenotype Differences between Human and Mouse Orthologous Genes.

Mice have been widely used as a model organism to investigate human gene-phenotype relationships based on a conjecture that orthologous genes generally perform similar functions and are associated with similar phenotypes. However, phenotypes associated with orthologous genes often turn out to be quite different between human and mouse. Herein, we devised a method to quantitatively compare phenotypes annotations associated with mouse models and human. Using semantic similarity comparisons, we identified orthologous genes with different phenotype annotations, of which the similarity score is on a par with that of random gene pairs. Analysis of sequence evolution and transcriptomic changes revealed that orthologous genes with phenotypic differences are correlated with changes in noncoding regulatory elements and tissue-specific expression profiles rather than changes in protein-coding sequences. To map accurate gene-phenotype relationships using model organisms, we propose that careful consideration of the evolutionary divergence of noncoding regulatory elements and transcriptomic profiles is essential.

Linear motif-mediated interactions have contributed to the evolution of modularity in complex protein interaction networks.

The modular architecture of protein-protein interaction (PPI) networks is evident in diverse species with a wide range of complexity. However, the molecular components that lead to the evolution of modularity in PPI networks have not been clearly identified. Here, we show that weak domain-linear motif interactions (DLIs) are more likely to connect different biological modules than strong domain-domain interactions (DDIs). This molecular division of labor is essential for the evolution of modularity in the complex PPI networks of diverse eukaryotic species. In particular, DLIs may compensate for the reduction in module boundaries that originate from increased connections between different modules in complex PPI networks. In addition, we show that the identification of biological modules can be greatly improved by including molecular characteristics of protein interactions. Our findings suggest that transient interactions have played a unique role in shaping the architecture and modularity of biological networks over the course of evolution.

Red2Flpe-SCON: a versatile, multicolor strategy for generating mosaic conditional knockout mice.

Image-based lineage tracing enables tissue turnover kinetics and lineage potentials of different adult cell populations to be investigated. Previously, we reported a genetic mouse model system, Red2Onco, which ectopically expressed mutated oncogenes together with red fluorescent proteins (RFP). This system enabled the expansion kinetics and neighboring effects of oncogenic clones to be dissected. We now report Red2Flpe-SCON: a mosaic knockout system that uses multicolor reporters to label both mutant and wild-type cells. We develop the Red2Flpe mouse line for red clone-specific Flpe expression, as well as the FRT-based SCON (Short Conditional IntrON) method to facilitate tunable conditional mosaic knockouts in mice. We use the Red2Flpe-SCON method to study Sox2 mutant clonal analysis in the esophageal epithelium of adult mice which reveal that the stem cell gene, Sox2, is less essential for adult stem cell maintenance itself, but rather for stem cell proliferation and differentiation.

Intestinal Paneth cell differentiation relies on asymmetric regulation of Wnt signaling by Daam1/2.

The mammalian intestine is one of the most rapidly self-renewing tissues, driven by stem cells residing at the crypt bottom. Paneth cells form a major element of the niche microenvironment providing various growth factors to orchestrate intestinal stem cell homeostasis, such as Wnt3. Different Wnt ligands can selectively activate ß-catenin-dependent (canonical) or -independent (noncanonical) signaling. Here, we report that the Dishevelled-associated activator of morphogenesis 1 (Daam1) and its paralogue Daam2 asymmetrically regulate canonical and noncanonical Wnt (Wnt/PCP) signaling. Daam1/2 interacts with the Wnt inhibitor RNF43, and Daam1/2 double knockout stimulates canonical Wnt signaling by preventing RNF43-dependent degradation of the Wnt receptor, Frizzled (Fzd). Single-cell RNA sequencing analysis revealed that Paneth cell differentiation is impaired by Daam1/2 depletion because of defective Wnt/PCP signaling. Together, we identified Daam1/2 as an unexpected hub molecule coordinating both canonical and noncanonical Wnt, which is fundamental for specifying an adequate number of Paneth cells.

SCON-a Short Conditional intrON for conditional knockout with one-step zygote injection.

The generation of conditional alleles using CRISPR technology is still challenging. Here, we introduce a Short Conditional intrON (SCON, 189 bp) that enables the rapid generation of conditional alleles via one-step zygote injection. In this study, a total of 13 SCON mouse lines were successfully generated by 2 different laboratories. SCON has conditional intronic functions in various vertebrate species, and its target insertion is as simple as CRISPR/Cas9-mediated gene tagging.

p57Kip2 imposes the reserve stem cell state of gastric chief cells.

Adult stem cells constantly react to local changes to ensure tissue homeostasis. In the main body of the stomach, chief cells produce digestive enzymes; however, upon injury, they undergo rapid proliferation for prompt tissue regeneration. Here, we identified p57Kip2 (p57) as a molecular switch for the reserve stem cell state of chief cells in mice. During homeostasis, p57 is constantly expressed in chief cells but rapidly diminishes after injury, followed by robust proliferation. Both single-cell RNA sequencing and dox-induced lineage tracing confirmed the sequential loss of p57 and activation of proliferation within the chief cell lineage. In corpus organoids, p57 overexpression induced a long-term reserve stem cell state, accompanied by altered niche requirements and a mature chief cell/secretory phenotype. Following the constitutive expression of p57 in vivo, chief cells showed an impaired injury response. Thus, p57 is a gatekeeper that imposes the reserve stem cell state of chief cells in homeostasis.

Domain-mediated interactions for protein subfamily identification.

Within a protein family, proteins with the same domain often exhibit different cellular functions, despite the shared evolutionary history and molecular function of the domain. We hypothesized that domain-mediated interactions (DMIs) may categorize a protein family into subfamilies because the diversified functions of a single domain often depend on interacting partners of domains. Here we systematically identified DMI subfamilies, in which proteins share domains with DMI partners, as well as with various functional and physical interaction networks in individual species. In humans, DMI subfamily members are associated with similar diseases, including cancers, and are frequently co-associated with the same diseases. DMI information relates to the functional and evolutionary subdivisions of human kinases. In yeast, DMI subfamilies contain proteins with similar phenotypic outcomes from specific chemical treatments. Therefore, the systematic investigation here provides insights into the diverse functions of subfamilies derived from a protein family with a link-centric approach and suggests a useful resource for annotating the functions and phenotypic outcomes of proteins.

Network-based machine learning in colorectal and bladder organoid models predicts anti-cancer drug efficacy in patients.

Cancer patient classification using predictive biomarkers for anti-cancer drug responses is essential for improving therapeutic outcomes. However, current machine-learning-based predictions of drug response often fail to identify robust translational biomarkers from preclinical models. Here, we present a machine-learning framework to identify robust drug biomarkers by taking advantage of network-based analyses using pharmacogenomic data derived from three-dimensional organoid culture models. The biomarkers identified by our approach accurately predict the drug responses of 114 colorectal cancer patients treated with 5-fluorouracil and 77 bladder cancer patients treated with cisplatin. We further confirm our biomarkers using external transcriptomic datasets of drug-sensitive and -resistant isogenic cancer cell lines. Finally, concordance analysis between the transcriptomic biomarkers and independent somatic mutation-based biomarkers further validate our method. This work presents a method to predict cancer patient drug responses using pharmacogenomic data derived from organoid models by combining the application of gene modules and network-based approaches.

Link clustering explains non-central and contextually essential genes in protein interaction networks.

Recent studies have shown that many essential genes (EGs) change their essentiality across various contexts. Finding contextual EGs in pathogenic conditions may facilitate the identification of therapeutic targets. We propose link clustering as an indicator of contextual EGs that are non-central in protein-protein interaction (PPI) networks. In various human and yeast PPI networks, we found that 29-47% of EGs were better characterized by link clustering than by centrality. Importantly, non-central EGs were prone to change their essentiality across different human cell lines and between species. Compared with central EGs and non-EGs, non-central EGs had intermediate levels of expression and evolutionary conservation. In addition, non-central EGs exhibited a significant impact on communities at lower hierarchical levels, suggesting that link clustering is associated with contextual essentiality, as it depicts locally important nodes in network structures.

Energy metabolism and whole-exome sequencing-based analysis of Sasang constitution: a pilot study.

BACKGROUND: Traditional Korean Sasang constitutional (SC) medicine categorizes individuals into four constitutional types [Tae-eum (TE), So-eum (SE), Tae-yang (TY), or So-yang (SY)] based on biological and physiological characteristics. As these characteristics are closely related to the bioenergetics of the human body, we assessed the correlation between SC type and energy metabolism features. METHODS: Forty healthy, young (22.3 ± 1.4 years) males volunteered to participate in this study. Participants answered an SC questionnaire, and their face shape, voice tone, and body shape were assessed using an SC analysis tool. Thirty-one participants (10 TE, 10 SE, 3 TY, and 8 SY) were selected for further analysis. Collected blood samples were subjected to blood composition analysis, mitochondrial function analysis, and whole-exome sequencing. RESULTS: The SY type showed significantly lower total cholesterol and high-density lipoprotein cholesterol levels than the SE type. Cellular and mitochondrial Adenosine triphosphate (ATP) levels were similar across types. All types showed similar basal mitochondrial oxygen consumption rates, whereas the TE type showed a significantly lower ATP-linked oxygen consumption rate than the other types. Whole-exome sequencing identified several genes variants that were exclusively detected in particular SC types, including 19 for SE, seven for SY, 11 for TE, and six for TY. CONCLUSION: SC type-specific differences in mitochondrial function and gene mutations were detected in a small group of healthy, young Korean males. These results are expected to greatly improve the accurate screening and utilization of SC medicine.

OASIS 2: online application for survival analysis 2 with features for the analysis of maximal lifespan and healthspan in aging research.

Online application for survival analysis (OASIS) has served as a popular and convenient platform for the statistical analysis of various survival data, particularly in the field of aging research. With the recent advances in the fields of aging research that deal with complex survival data, we noticed a need for updates to the current version of OASIS. Here, we report OASIS 2 (http://sbi.postech.ac.kr/oasis2), which provides extended statistical tools for survival data and an enhanced user interface. In particular, OASIS 2 enables the statistical comparison of maximal lifespans, which is potentially useful for determining key factors that limit the lifespan of a population. Furthermore, OASIS 2 provides statistical and graphical tools that compare values in different conditions and times. That feature is useful for comparing age-associated changes in physiological activities, which can be used as indicators of "healthspan." We believe that OASIS 2 will serve as a standard platform for survival analysis with advanced and user-friendly statistical tools for experimental biologists in the field of aging research.