Improving bulk RNA-seq classification by transferring gene signature from single cells in acute myeloid leukemia.

The advances in single-cell RNA sequencing (scRNA-seq) technologies enable the characterization of transcriptomic profiles at the cellular level and demonstrate great promise in bulk sample analysis thereby offering opportunities to transfer gene signature from scRNA-seq to bulk data. However, the gene expression signatures identified from single cells are typically inapplicable to bulk RNA-seq data due to the profiling differences of distinct sequencing technologies. Here, we propose single-cell pair-wise gene expression (scPAGE), a novel method to develop single-cell gene pair signatures (scGPSs) that were beneficial to bulk RNA-seq classification to transfer knowledge across platforms. PAGE was adopted to tackle the challenge of profiling differences. We applied the method to acute myeloid leukemia (AML) and identified the scGPS from mouse scRNA-seq that allowed discriminating between AML and control cells. The scGPS was validated in bulk RNA-seq datasets and demonstrated better performance (average area under the curve [AUC] = 0.96) than the conventional gene expression strategies (average AUC$\le$ 0.88) suggesting its potential in disclosing the molecular mechanism of AML. The scGPS also outperformed its bulk counterpart, which highlighted the benefit of gene signature transfer. Furthermore, we confirmed the utility of scPAGE in sepsis as an example of other disease scenarios. scPAGE leveraged the advantages of single-cell profiles to enhance the analysis of bulk samples revealing great potential of transferring knowledge from single-cell to bulk transcriptome studies.

SHARP: hyperfast and accurate processing of single-cell RNA-seq data via ensemble random projection.

To process large-scale single-cell RNA-sequencing (scRNA-seq) data effectively without excessive distortion during dimension reduction, we present SHARP, an ensemble random projection-based algorithm that is scalable to clustering 10 million cells. Comprehensive benchmarking tests on 17 public scRNA-seq data sets show that SHARP outperforms existing methods in terms of speed and accuracy. Particularly, for large-size data sets (more than 40,000 cells), SHARP runs faster than other competitors while maintaining high clustering accuracy and robustness. To the best of our knowledge, SHARP is the only R-based tool that is scalable to clustering scRNA-seq data with 10 million cells.

Genomic profiling identifies genes and pathways dysregulated by HEY1-NCOA2 fusion and shines a light on mesenchymal chondrosarcoma tumorigenesis.

Mesenchymal chondrosarcoma is a rare, high-grade, primitive mesenchymal tumor. It accounts for around 2-10% of all chondrosarcomas and mainly affects adolescents and young adults. We previously described the HEY1-NCOA2 as a recurrent gene fusion in mesenchymal chondrosarcoma, an important breakthrough for characterizing this disease; however, little study had been done to characterize the fusion protein functionally, in large part due to a lack of suitable models for evaluating the impact of HEY1-NCOA2 expression in the appropriate cellular context. We used iPSC-derived mesenchymal stem cells (iPSC-MSCs), which can differentiate into chondrocytes, and generated stable transduced iPSC-MSCs with inducible expression of HEY1-NCOA2 fusion protein, wildtype HEY1 or wildtype NCOA2. We next comprehensively analyzed both the DNA binding properties and transcriptional impact of HEY1-NCOA2 expression by integrating genome-wide chromatin immunoprecipitation sequencing (ChIP-seq) and expression profiling (RNA-seq). We demonstrated that HEY1-NCOA2 fusion protein preferentially binds to promoter regions of canonical HEY1 targets, resulting in transactivation of HEY1 targets, and significantly enhances cell proliferation. Intriguingly, we identified that both PDGFB and PDGFRA were directly targeted and upregulated by HEY1-NCOA2; and the fusion protein, but not wildtype HEY1 or NCOA2, dramatically increased the level of phospho-AKT (Ser473). Our findings provide a rationale for exploring PDGF/PI3K/AKT inhibition in treating mesenchymal chondrosarcoma. © 2022 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

YAP/TAZ maintain the proliferative capacity and structural organization of radial glial cells during brain development.

The Hippo pathway regulates the development and homeostasis of many tissues and in many species. It controls the activity of two paralogous transcriptional coactivators, YAP and TAZ (YAP/TAZ). Although previous studies have established that aberrant YAP/TAZ activation is detrimental to mammalian brain development, whether and how endogenous levels of YAP/TAZ activity regulate brain development remain unclear. Here, we show that during mammalian cortical development, YAP/TAZ are specifically expressed in apical neural progenitor cells known as radial glial cells (RGCs). The subcellular localization of YAP/TAZ undergoes dynamic changes as corticogenesis proceeds. YAP/TAZ are required for maintaining the proliferative potential and structural organization of RGCs, and their ablation during cortical development reduces the numbers of cortical projection neurons and causes the loss of ependymal cells, resulting in hydrocephaly. Transcriptomic analysis using sorted RGCs reveals gene expression changes in YAP/TAZ-depleted cells that correlate with mutant phenotypes. Thus, our study has uncovered essential functions of YAP/TAZ during mammalian brain development and revealed the transcriptional mechanism of their action.

A Critical Role for Estrogen-Related Receptor Signaling in Cardiac Maturation.

RATIONALE: The heart undergoes dramatic developmental changes during the prenatal to postnatal transition, including maturation of cardiac myocyte energy metabolic and contractile machinery. Delineation of the mechanisms involved in cardiac postnatal development could provide new insight into the fetal shifts that occur in the diseased heart and unveil strategies for driving maturation of stem cell-derived cardiac myocytes. OBJECTIVE: To delineate transcriptional drivers of cardiac maturation. METHODS AND RESULTS: We hypothesized that ERR (estrogen-related receptor) α and γ, known transcriptional regulators of postnatal mitochondrial biogenesis and function, serve a role in the broader cardiac maturation program. We devised a strategy to knockdown the expression of ERRα and γ in heart after birth (pn-csERRα/γ [postnatal cardiac-specific ERRα/γ]) in mice. With high levels of knockdown, pn-csERRα/γ knockdown mice exhibited cardiomyopathy with an arrest in mitochondrial maturation. RNA sequence analysis of pn-csERRα/γ knockdown hearts at 5 weeks of age combined with chromatin immunoprecipitation with deep sequencing and functional characterization conducted in human induced pluripotent stem cell-derived cardiac myocytes (hiPSC-CM) demonstrated that ERRγ activates transcription of genes involved in virtually all aspects of postnatal developmental maturation, including mitochondrial energy transduction, contractile function, and ion transport. In addition, ERRγ was found to suppress genes involved in fibroblast activation in hearts of pn-csERRα/γ knockdown mice. Disruption of Esrra and Esrrg in mice during fetal development resulted in perinatal lethality associated with structural and genomic evidence of an arrest in cardiac maturation, including persistent expression of early developmental and noncardiac lineage gene markers including cardiac fibroblast signatures. Lastly, targeted deletion of ESRRA and ESRRG in hiPSC-CM derepressed expression of early (transcription factor 21 or TCF21) and mature (periostin, collagen type III) fibroblast gene signatures. CONCLUSIONS: ERRα and γ are critical regulators of cardiac myocyte maturation, serving as transcriptional activators of adult cardiac metabolic and structural genes, an.d suppressors of noncardiac lineages including fibroblast determination.

FUEL-mLoc: feature-unified prediction and explanation of multi-localization of cellular proteins in multiple organisms.

Although many web-servers for predicting protein subcellular localization have been developed, they often have the following drawbacks: (i) lack of interpretability or interpreting results with heterogenous information which may confuse users; (ii) ignoring multi-location proteins and (iii) only focusing on specific organism. To tackle these problems, we present an interpretable and efficient web-server, namely FUEL-mLoc, using eature- nified prediction and xplanation of m ulti- oc alization of cellular proteins in multiple organisms. Compared to conventional localization predictors, FUEL-mLoc has the following advantages: (i) using unified features (i.e. essential GO terms) to interpret why a prediction is made; (ii) being capable of predicting both single- and multi-location proteins and (iii) being able to handle proteins of multiple organisms, including Eukaryota, Homo sapiens, Viridiplantae, Gram-positive Bacteria, Gram-negative Bacteria and Virus . Experimental results demonstrate that FUEL-mLoc outperforms state-of-the-art subcellular-localization predictors. Availability and Implementation: http://bioinfo.eie.polyu.edu.hk/FUEL-mLoc/. Contacts: shibiao.wan@princeton.edu or enmwmak@polyu.edu.hk. Supplementary information: Supplementary data are available at Bioinformatics online.

Sparse regressions for predicting and interpreting subcellular localization of multi-label proteins.

BACKGROUND: Predicting protein subcellular localization is indispensable for inferring protein functions. Recent studies have been focusing on predicting not only single-location proteins, but also multi-location proteins. Almost all of the high performing predictors proposed recently use gene ontology (GO) terms to construct feature vectors for classification. Despite their high performance, their prediction decisions are difficult to interpret because of the large number of GO terms involved. RESULTS: This paper proposes using sparse regressions to exploit GO information for both predicting and interpreting subcellular localization of single- and multi-location proteins. Specifically, we compared two multi-label sparse regression algorithms, namely multi-label LASSO (mLASSO) and multi-label elastic net (mEN), for large-scale predictions of protein subcellular localization. Both algorithms can yield sparse and interpretable solutions. By using the one-vs-rest strategy, mLASSO and mEN identified 87 and 429 out of more than 8,000 GO terms, respectively, which play essential roles in determining subcellular localization. More interestingly, many of the GO terms selected by mEN are from the biological process and molecular function categories, suggesting that the GO terms of these categories also play vital roles in the prediction. With these essential GO terms, not only where a protein locates can be decided, but also why it resides there can be revealed. CONCLUSIONS: Experimental results show that the output of both mEN and mLASSO are interpretable and they perform significantly better than existing state-of-the-art predictors. Moreover, mEN selects more features and performs better than mLASSO on a stringent human benchmark dataset. For readers' convenience, an online server called SpaPredictor for both mLASSO and mEN is available at http://bioinfo.eie.polyu.edu.hk/SpaPredictorServer/.

Ensemble Linear Neighborhood Propagation for Predicting Subchloroplast Localization of Multi-Location Proteins.

In the postgenomic era, the number of unreviewed protein sequences is remarkably larger and grows tremendously faster than that of reviewed ones. However, existing methods for protein subchloroplast localization often ignore the information from these unlabeled proteins. This paper proposes a multi-label predictor based on ensemble linear neighborhood propagation (LNP), namely, LNP-Chlo, which leverages hybrid sequence-based feature information from both labeled and unlabeled proteins for predicting localization of both single- and multi-label chloroplast proteins. Experimental results on a stringent benchmark dataset and a novel independent dataset suggest that LNP-Chlo performs at least 6% (absolute) better than state-of-the-art predictors. This paper also demonstrates that ensemble LNP significantly outperforms LNP based on individual features. For readers' convenience, the online Web server LNP-Chlo is freely available at http://bioinfo.eie.polyu.edu.hk/LNPChloServer/ .

Mem-ADSVM: A two-layer multi-label predictor for identifying multi-functional types of membrane proteins.

Identifying membrane proteins and their multi-functional types is an indispensable yet challenging topic in proteomics and bioinformatics. However, most of the existing membrane-protein predictors have the following problems: (1) they do not predict whether a given protein is a membrane protein or not; (2) they are limited to predicting membrane proteins with single-label functional types but ignore those with multi-functional types; and (3) there is still much room for improvement for their performance. To address these problems, this paper proposes a two-layer multi-label predictor, namely Mem-ADSVM, which can identify membrane proteins (Layer I) and their multi-functional types (Layer II). Specifically, given a query protein, its associated gene ontology (GO) information is retrieved by searching a compact GO-term database with its homologous accession number. Subsequently, the GO information is classified by a binary support vector machine (SVM) classifier to determine whether it is a membrane protein or not. If yes, it will be further classified by a multi-label multi-class SVM classifier equipped with an adaptive-decision (AD) scheme to determine to which functional type(s) it belongs. Experimental results show that Mem-ADSVM significantly outperforms state-of-the-art predictors in terms of identifying both membrane proteins and their multi-functional types. This paper also suggests that the two-layer prediction architecture is better than the one-layer for prediction performance. For reader׳s convenience, the Mem-ADSVM server is available online at http://bioinfo.eie.polyu.edu.hk/MemADSVMServer/.

mPLR-Loc: an adaptive decision multi-label classifier based on penalized logistic regression for protein subcellular localization prediction.

Proteins located in appropriate cellular compartments are of paramount importance to exert their biological functions. Prediction of protein subcellular localization by computational methods is required in the post-genomic era. Recent studies have been focusing on predicting not only single-location proteins but also multi-location proteins. However, most of the existing predictors are far from effective for tackling the challenges of multi-label proteins. This article proposes an efficient multi-label predictor, namely mPLR-Loc, based on penalized logistic regression and adaptive decisions for predicting both single- and multi-location proteins. Specifically, for each query protein, mPLR-Loc exploits the information from the Gene Ontology (GO) database by using its accession number (AC) or the ACs of its homologs obtained via BLAST. The frequencies of GO occurrences are used to construct feature vectors, which are then classified by an adaptive decision-based multi-label penalized logistic regression classifier. Experimental results based on two recent stringent benchmark datasets (virus and plant) show that mPLR-Loc remarkably outperforms existing state-of-the-art multi-label predictors. In addition to being able to rapidly and accurately predict subcellular localization of single- and multi-label proteins, mPLR-Loc can also provide probabilistic confidence scores for the prediction decisions. For readers' convenience, the mPLR-Loc server is available online (http://bioinfo.eie.polyu.edu.hk/mPLRLocServer).

mLASSO-Hum: A LASSO-based interpretable human-protein subcellular localization predictor.

Knowing the subcellular compartments of human proteins is essential to shed light on the mechanisms of a broad range of human diseases. In computational methods for protein subcellular localization, knowledge-based methods (especially gene ontology (GO) based methods) are known to perform better than sequence-based methods. However, existing GO-based predictors often lack interpretability and suffer from overfitting due to the high dimensionality of feature vectors. To address these problems, this paper proposes an interpretable multi-label predictor, namely mLASSO-Hum, which can yield sparse and interpretable solutions for large-scale prediction of human protein subcellular localization. By using the one-vs-rest LASSO-based classifiers, 87 out of more than 8000 GO terms are found to play more significant roles in determining the subcellular localization. Based on these 87 essential GO terms, we can decide not only where a protein resides within a cell, but also why it is located there. To further exploit information from the remaining GO terms, a method based on the GO hierarchical information derived from the depth distance of GO terms is proposed. Experimental results show that mLASSO-Hum performs significantly better than state-of-the-art predictors. We also found that in addition to the GO terms from the cellular component category, GO terms from the other two categories also play important roles in the final classification decisions. For readers' convenience, the mLASSO-Hum server is available online at http://bioinfo.eie.polyu.edu.hk/mLASSOHumServer/.

R3P-Loc: a compact multi-label predictor using ridge regression and random projection for protein subcellular localization.

Locating proteins within cellular contexts is of paramount significance in elucidating their biological functions. Computational methods based on knowledge databases (such as gene ontology annotation (GOA) database) are known to be more efficient than sequence-based methods. However, the predominant scenarios of knowledge-based methods are that (1) knowledge databases typically have enormous size and are growing exponentially, (2) knowledge databases contain redundant information, and (3) the number of extracted features from knowledge databases is much larger than the number of data samples with ground-truth labels. These properties render the extracted features liable to redundant or irrelevant information, causing the prediction systems suffer from overfitting. To address these problems, this paper proposes an efficient multi-label predictor, namely R3P-Loc, which uses two compact databases for feature extraction and applies random projection (RP) to reduce the feature dimensions of an ensemble ridge regression (RR) classifier. Two new compact databases are created from Swiss-Prot and GOA databases. These databases possess almost the same amount of information as their full-size counterparts but with much smaller size. Experimental results on two recent datasets (eukaryote and plant) suggest that R3P-Loc can reduce the dimensions by seven-folds and significantly outperforms state-of-the-art predictors. This paper also demonstrates that the compact databases reduce the memory consumption by 39 times without causing degradation in prediction accuracy. For readers׳ convenience, the R3P-Loc server is available online at url:http://bioinfo.eie.polyu.edu.hk/R3PLocServer/.

A Review for Artificial Intelligence Based Protein Subcellular Localization.

Proteins need to be located in appropriate spatiotemporal contexts to carry out their diverse biological functions. Mislocalized proteins may lead to a broad range of diseases, such as cancer and Alzheimer's disease. Knowing where a target protein resides within a cell will give insights into tailored drug design for a disease. As the gold validation standard, the conventional wet lab uses fluorescent microscopy imaging, immunoelectron microscopy, and fluorescent biomarker tags for protein subcellular location identification. However, the booming era of proteomics and high-throughput sequencing generates tons of newly discovered proteins, making protein subcellular localization by wet-lab experiments a mission impossible. To tackle this concern, in the past decades, artificial intelligence (AI) and machine learning (ML), especially deep learning methods, have made significant progress in this research area. In this article, we review the latest advances in AI-based method development in three typical types of approaches, including sequence-based, knowledge-based, and image-based methods. We also elaborately discuss existing challenges and future directions in AI-based method development in this research field.

SAMP: Identifying Antimicrobial Peptides by an Ensemble Learning Model Based on Proportionalized Split Amino Acid Composition.

It is projected that 10 million deaths could be attributed to drug-resistant bacteria infections in 2050. To address this concern, identifying new-generation antibiotics is an effective way. Antimicrobial peptides (AMPs), a class of innate immune effectors, have received significant attention for their capacity to eliminate drug-resistant pathogens, including viruses, bacteria, and fungi. Recent years have witnessed widespread applications of computational methods especially machine learning (ML) and deep learning (DL) for discovering AMPs. However, existing methods only use features including compositional, physiochemical, and structural properties of peptides, which cannot fully capture sequence information from AMPs. Here, we present SAMP, an ensemble random projection (RP) based computational model that leverages a new type of features called Proportionalized Split Amino Acid Composition (PSAAC) in addition to conventional sequence-based features for AMP prediction. With this new feature set, SAMP captures the residue patterns like sorting signals at around both the N-terminus and the C-terminus, while also retaining the sequence order information from the middle peptide fragments. Benchmarking tests on different balanced and imbalanced datasets demonstrate that SAMP consistently outperforms existing state-of-the-art methods, such as iAMPpred and AMPScanner V2, in terms of accuracy, MCC, G-measure and F1-score. In addition, by leveraging an ensemble RP architecture, SAMP is scalable to processing large-scale AMP identification with further performance improvement, compared to those models without RP. To facilitate the use of SAMP, we have developed a Python package freely available at https://github.com/wan-mlab/SAMP.

The context-dependent epigenetic and organogenesis programs determine 3D vs. 2D cellular fitness of MYC-driven cancer.

3D cellular-specific epigenetic and transcriptomic reprogramming is critical to organogenesis and tumorigenesis. Here we dissect the distinct cell fitness in 2D (normoxia vs. chronic hypoxia) vs 3D (normoxia) culture conditions. We identify over 600 shared essential genes and additional context-specific fitness genes and pathways. Knockout of the VHL-HIF1 pathway results in incompatible fitness defects under normoxia vs. 1% oxygen or 3D culture conditions. Moreover, deletion of each of the mitochondrial respiratory electron transport chain complex has distinct fitness outcomes. Notably, multicellular organogenesis signaling pathways including TGFß-SMAD specifically constrict the uncontrolled cell proliferation in 3D while inactivation of epigenetic modifiers (Bcor, Kmt2d, Mettl3 and Mettl14) has opposite outcomes in 2D vs. 3D. We further identify a 3D-dependent synthetic lethality with partial loss of Prmt5 due to a reduction of Mtap expression resulting from 3D-specific epigenetic reprogramming. Our study highlights unique epigenetic, metabolic and organogenesis signaling dependencies under different cellular settings.

GOASVM: a subcellular location predictor by incorporating term-frequency gene ontology into the general form of Chou's pseudo-amino acid composition.

Prediction of protein subcellular localization is an important yet challenging problem. Recently, several computational methods based on Gene Ontology (GO) have been proposed to tackle this problem and have demonstrated superiority over methods based on other features. Existing GO-based methods, however, do not fully use the GO information. This paper proposes an efficient GO method called GOASVM that exploits the information from the GO term frequencies and distant homologs to represent a protein in the general form of Chou's pseudo-amino acid composition. The method first selects a subset of relevant GO terms to form a GO vector space. Then for each protein, the method uses the accession number (AC) of the protein or the ACs of its homologs to find the number of occurrences of the selected GO terms in the Gene Ontology annotation (GOA) database as a means to construct GO vectors for support vector machines (SVMs) classification. With the advantages of GO term frequencies and a new strategy to incorporate useful homologous information, GOASVM can achieve a prediction accuracy of 72.2% on a new independent test set comprising novel proteins that were added to Swiss-Prot six years later than the creation date of the training set. GOASVM and Supplementary materials are available online at http://bioinfo.eie.polyu.edu.hk/mGoaSvmServer/GOASVM.html.

Embedded Bioprinting of Breast Tumor Cells and Organoids Using Low-Concentration Collagen-Based Bioinks.

Bioinks for 3D bioprinting of tumor models should not only meet printability requirements but also accurately maintain and support phenotypes of tumor surrounding cells to recapitulate key tumor hallmarks. Collagen is a major extracellular matrix protein for solid tumors, but low viscosity of collagen solution has made 3D bioprinted cancer models challenging. This work produces embedded, bioprinted breast cancer cells and tumor organoid models using low-concentration collagen I based bioinks. The biocompatible and physically crosslinked silk fibroin hydrogel is used to generate the support bath for the embedded 3D printing. The composition of the collagen I based bioink is optimized with a thermoresponsive hyaluronic acid-based polymer to maintain the phenotypes of both the noninvasive epithelial and invasive breast cancer cells, as well as cancer-associated fibroblasts. Mouse breast tumor organoids are bioprinted using optimized collagen bioink to mimic in vivo tumor morphology. A vascularized tumor model is also created using a similar strategy, with significantly enhanced vasculature formation under hypoxia. This study shows the great potential of embedded bioprinted breast tumor models utilizing a low-concentration collagen-based bioink for advancing the understanding of tumor cell biology and facilitating drug discovery research.

Procyanidin alleviates ferroptosis and inflammation of LPS-induced RAW264.7 cell via the Nrf2/HO-1 pathway.

Inflammation is a common occurrence in many medical conditions and is a natural defense mechanism of the human body. Ferroptosis, an iron-dependent form of cell death related to lipid peroxide build-up, has been found to be involved in inflammation. The anti-inflammatory effects of procyanidin, however, are not yet fully understood. Through network pharmacology and bioinformatics analysis, it was suggested that procyanidin could modulate ferroptosis and cause anti-inflammatory effects on RAW264.7 cells. This was further evidenced through molecular docking, molecular dynamics, and in vitro experiments. The results indicated that procyanidin could diminish inflammation in LPS-induced RAW264.7 cells by regulating ferroptosis via the Nrf2/HO-1/Keap-1 pathway. In conclusion, procyanidin supplementation might be an effective way to reduce inflammation by decreasing the release of inflammatory cytokines and suppressing ferroptosis.

USP1 Expression Driven by EWS::FLI1 Transcription Factor Stabilizes Survivin and Mitigates Replication Stress in Ewing Sarcoma.

In this study, we identify USP1 as a transcriptional target of EWS::FLI1 and demonstrate the requisite function of USP1 in Ewing sarcoma (EWS) cell survival in response to endogenous replication stress. EWS::FLI1 oncogenic transcription factor drives most EWS, a pediatric bone cancer. EWS cells display elevated levels of R-loops and replication stress. The mechanism by which EWS cells override activation of apoptosis or cellular senescence in response to increased replication stress is not known. We show that USP1 is overexpressed in EWS and EWS::FLI1 regulates USP1 transcript levels. USP1 knockdown or inhibition arrests EWS cell growth and induces cell death by apoptosis. Mechanistically, USP1 regulates Survivin (BIRC5/API4) protein stability and the activation of caspase-9 and caspase-3/7 in response to endogenous replication stress. Notably, USP1 inhibition sensitizes cells to doxorubicin and etoposide treatment. Together, our study demonstrates that USP1 is regulated by EWS::FLI1, the USP1-Survivin axis promotes EWS cell survival, and USP1 inhibition sensitizes cells to standard of care chemotherapy. IMPLICATIONS: High USP1 and replication stress levels driven by EWS::FLI1 transcription factor in EWS are vulnerabilities that can be exploited to improve existing treatment avenues and overcome drug resistance.

Etiology of oncogenic fusions in 5,190 childhood cancers and its clinical and therapeutic implication.

Oncogenic fusions formed through chromosomal rearrangements are hallmarks of childhood cancer that define cancer subtype, predict outcome, persist through treatment, and can be ideal therapeutic targets. However, mechanistic understanding of the etiology of oncogenic fusions remains elusive. Here we report a comprehensive detection of 272 oncogenic fusion gene pairs by using tumor transcriptome sequencing data from 5190 childhood cancer patients. We identify diverse factors, including translation frame, protein domain, splicing, and gene length, that shape the formation of oncogenic fusions. Our mathematical modeling reveals a strong link between differential selection pressure and clinical outcome in CBFB-MYH11. We discover 4 oncogenic fusions, including RUNX1-RUNX1T1, TCF3-PBX1, CBFA2T3-GLIS2, and KMT2A-AFDN, with promoter-hijacking-like features that may offer alternative strategies for therapeutic targeting. We uncover extensive alternative splicing in oncogenic fusions including KMT2A-MLLT3, KMT2A-MLLT10, C11orf95-RELA, NUP98-NSD1, KMT2A-AFDN and ETV6-RUNX1. We discover neo splice sites in 18 oncogenic fusion gene pairs and demonstrate that such splice sites confer therapeutic vulnerability for etiology-based genome editing. Our study reveals general principles on the etiology of oncogenic fusions in childhood cancer and suggests profound clinical implications including etiology-based risk stratification and genome-editing-based therapeutics.