FusionGDB 2.0: fusion gene annotation updates aided by deep learning

Artificial intelligence in fusion protein three-dimensional structure prediction: Review and perspective

Recent advancements in artificial intelligence (AI) have accelerated the prediction of unknown protein structures. However, accurately predicting the three-dimensional (3D) structures of fusion proteins remains a difficult task because the current AI-based protein structure predictions are focused on the WT proteins rather than on the newly fused proteins in nature. Following the central dogma of biology, fusion proteins are translated from fusion transcripts, which are made by transcribing the fusion genes between two different loci through the chromosomal rearrangements in cancer. Accurately predicting the 3D structures of fusion proteins is important for understanding the functional roles and mechanisms of action of new chimeric proteins. However, predicting their 3D structure using a template-based model is challenging because known template structures are often unavailable in databases. Deep learning (DL) models that utilize multi-level protein information have revolutionized the prediction of protein 3D structures. In this review paper, we highlighted the latest advancements and ongoing challenges in predicting the 3D structure of fusion proteins using DL models. We aim to explore both the advantages and challenges of employing AlphaFold2, RoseTTAFold, tr-Rosetta and D-I-TASSER for modelling the 3D structures. HIGHLIGHTS: This review provides the overall pipeline and landscape of the prediction of the 3D structure of fusion protein. This review provides the factors that should be considered in predicting the 3D structures of fusion proteins using AI approaches in each step. This review highlights the latest advancements and ongoing challenges in predicting the 3D structure of fusion proteins using deep learning models. This review explores the advantages and challenges of employing AlphaFold2, RoseTTAFold, tr-Rosetta, and D-I-TASSER to model 3D structures.

Transcriptome-wide association analysis identifies candidate susceptibility genes for prostate-specific antigen levels in men without prostate cancer

Deciphering the genetic basis of prostate-specific antigen (PSA) levels may improve their utility for prostate cancer (PCa) screening. Using genome-wide association study (GWAS) summary statistics from 95,768 PCa-free men, we conducted a transcriptome-wide association study (TWAS) to examine impacts of genetically predicted gene expression on PSA. Analyses identified 41 statistically significant (p < 0.05/12,192 = 4.10 × 10-6) associations in whole blood and 39 statistically significant (p < 0.05/13,844 = 3.61 × 10-6) associations in prostate tissue, with 18 genes associated in both tissues. Cross-tissue analyses identified 155 statistically significantly (p < 0.05/22,249 = 2.25 × 10-6) genes. Out of 173 unique PSA-associated genes across analyses, we replicated 151 (87.3%) in a TWAS of 209,318 PCa-free individuals from the Million Veteran Program. Based on conditional analyses, we found 20 genes (11 single tissue, nine cross-tissue) that were associated with PSA levels in the discovery TWAS that were not attributable to a lead variant from a GWAS. Ten of these 20 genes replicated, and two of the replicated genes had colocalization probability of >0.5: CCNA2 and HIST1H2BN. Six of the 20 identified genes are not known to impact PCa risk. Fine-mapping based on whole blood and prostate tissue revealed five protein-coding genes with evidence of causal relationships with PSA levels. Of these five genes, four exhibited evidence of colocalization and one was conditionally independent of previous GWAS findings. These results yield hypotheses that should be further explored to improve understanding of genetic factors underlying PSA levels.

Transformers meets neoantigen detection: a systematic literature review

Cancer immunology offers a new alternative to traditional cancer treatments, such as radiotherapy and chemotherapy. One notable alternative is the development of personalized vaccines based on cancer neoantigens. Moreover, Transformers are considered a revolutionary development in artificial intelligence with a significant impact on natural language processing (NLP) tasks and have been utilized in proteomics studies in recent years. In this context, we conducted a systematic literature review to investigate how Transformers are applied in each stage of the neoantigen detection process. Additionally, we mapped current pipelines and examined the results of clinical trials involving cancer vaccines.

Recurrent and novel fusions detected by targeted RNA sequencing as part of the diagnostic workflow of soft tissue and bone tumours

The identification of gene fusions has become an integral part of soft tissue and bone tumour diagnosis. We investigated the added value of targeted RNA-based sequencing (targeted RNA-seq, Archer FusionPlex) to our current molecular diagnostic workflow of these tumours, which is based on fluorescence in situ hybridisation (FISH) for the detection of gene fusions using 25 probes. In a series of 131 diagnostic samples targeted RNA-seq identified a gene fusion, BCOR internal tandem duplication or ALK deletion in 47 cases (35.9%). For 74 cases, encompassing 137 FISH analyses, concordance between FISH and targeted RNA-seq was evaluated. A positive or negative FISH result was confirmed by targeted RNA-seq in 27 out of 49 (55.1%) and 81 out of 88 (92.0%) analyses, respectively. While negative concordance was high, targeted RNA-seq identified a canonical gene fusion in seven cases despite a negative FISH result. The 22 discordant FISH-positive analyses showed a lower percentage of rearrangement-positive nuclei (range 15-41%) compared to the concordant FISH-positive analyses (>41% of nuclei in 88.9% of cases). Six FISH analyses (in four cases) were finally considered false positive based on histological and targeted RNA-seq findings. For the EWSR1 FISH probe, we observed a gene-dependent disparity (p = 0.0020), with 8 out of 35 cases showing a discordance between FISH and targeted RNA-seq (22.9%). This study demonstrates an added value of targeted RNA-seq to our current diagnostic workflow of soft tissue and bone tumours in 19 out of 131 cases (14.5%), which we categorised as altered diagnosis (3 cases), added precision (6 cases), or augmented spectrum (10 cases). In the latter subgroup, four novel fusion transcripts were found for which the clinical relevance remains unclear: NAB2::NCOA2, YAP1::NUTM2B, HSPA8::BRAF, and PDE2A::PLAG1. Overall, targeted RNA-seq has proven extremely valuable in the diagnostic workflow of soft tissue and bone tumours.

FusionNW, a potential clinical impact assessment of kinases in pan-cancer fusion gene network

Kinase fusion genes are the most active fusion gene group in human cancer fusion genes. To help choose the clinically significant kinase so that the cancer patients that have fusion genes can be better diagnosed, we need a metric to infer the assessment of kinases in pan-cancer fusion genes rather than relying on the sample frequency expressed fusion genes. Most of all, multiple studies assessed human kinases as the drug targets using multiple types of genomic and clinical information, but none used the kinase fusion genes in their study. The assessment studies of kinase without kinase fusion gene events can miss the effect of one of the mechanisms that enhance the kinase function in cancer. To fill this gap, in this study, we suggest a novel way of assessing genes using a network propagation approach to infer how likely individual kinases influence the kinase fusion gene network composed of ~5K kinase fusion gene pairs. To select a better seed of propagation, we chose the top genes via dimensionality reduction like a principal component or latent layer information of six features of individual genes in pan-cancer fusion genes. Our approach may provide a novel way to assess of human kinases in cancer.

FusionNeoAntigen: a resource of fusion gene-specific neoantigens

Among the diverse sources of neoantigens (i.e. single-nucleotide variants (SNVs), insertions or deletions (Indels) and fusion genes), fusion gene-derived neoantigens are generally more immunogenic, have multiple targets per mutation and are more widely distributed across various cancer types. Therefore, fusion gene-derived neoantigens are a potential source of highly immunogenic neoantigens and hold great promise for cancer immunotherapy. However, the lack of fusion protein sequence resources and knowledge prevents this application. We introduce 'FusionNeoAntigen', a dedicated resource for fusion-specific neoantigens, accessible at https://compbio.uth.edu/FusionNeoAntigen. In this resource, we provide fusion gene breakpoint crossing neoantigens focused on ∼43K fusion proteins of ∼16K in-frame fusion genes from FusionGDB2.0. FusionNeoAntigen provides fusion gene information, corresponding fusion protein sequences, fusion breakpoint peptide sequences, fusion gene-derived neoantigen prediction, virtual screening between fusion breakpoint peptides having potential fusion neoantigens and human leucocyte antigens (HLAs), fusion breakpoint RNA/protein sequences for developing vaccines, information on samples with fusion-specific neoantigen, potential CAR-T targetable cell-surface fusion proteins and literature curation. FusionNeoAntigen will help to develop fusion gene-based immunotherapies. We will report all potential fusion-specific neoantigens from all possible open reading frames of ∼120K human fusion genes in future versions.

FusionPDB: a knowledgebase of human fusion proteins

Tumorigenic functions due to the formation of fusion genes have been targeted for cancer therapeutics (i.e. kinase inhibitors). However, many fusion proteins involved in various cellular processes have not been studied for targeted therapeutics. This is because the lack of complete fusion protein sequences and their whole 3D structures has made it challenging to develop new therapeutic strategies. To fill these critical gaps, we developed a computational pipeline and a resource of human fusion proteins named FusionPDB, available at https://compbio.uth.edu/FusionPDB. FusionPDB is organized into four levels: 43K fusion protein sequences (14.7K in-frame fusion genes, Level 1), over 2300 + 1267 fusion protein 3D structures (from 2300 recurrent and 266 manually curated in-frame fusion genes, Level 2), pLDDT score analysis for the 1267 fusion proteins from 266 manually curated fusion genes (Level 3), and virtual screening outcomes for 68 selected fusion proteins from 266 manually curated fusion genes (Level 4). FusionPDB is the only resource providing whole 3D structures of fusion proteins and comprehensive knowledge of human fusion proteins. It will be regularly updated until it covers all human fusion proteins in the future.

Systematic investigation of the homology sequences around the human fusion gene breakpoints in pan-cancer - bioinformatics study for a potential link to MMEJ

Microhomology-mediated end joining (MMEJ), an error-prone DNA damage repair mechanism, frequently leads to chromosomal rearrangements due to its ability to engage in promiscuous end joining of genomic instability and also leads to increasing mutational load at the sequences flanking the breakpoints (BPs). In this study, we systematically investigated the homology sequences around the genomic breakpoint area of human fusion genes, which were formed by the chromosomal rearrangements initiated by DNA double-strand breakage. Since the RNA-seq data is the typical data set to check the fusion genes, for the known exon junction fusion breakpoints identified from RNA-seq data, we have to infer the high chance of genomic breakpoint regions. For this, we utilized the high feature importance score area calculated from our recently developed fusion BP prediction model, FusionAI and identified 151 K microhomologies among ~24 K fusion BPs in 20 K fusion genes. From our multiple bioinformatics studies, we found a relationship between sequence homologies and the immune system. This in-silico study will provide novel knowledge on the sequence homologies around the coded structural variants.

Diffusion-enhanced characterization of 3D chromatin structure reveals its linkage to gene regulatory networks and the interactome

The interactome networks at the DNA, RNA, and protein levels are crucial for cellular functions, and the diverse variations of these networks are heavily involved in the establishment of different cell states. We have developed a diffusion-based method, Hi-C to geometry (CTG), to obtain reliable geometric information on the chromatin from Hi-C data. CTG produces a consistent and reproducible framework for the 3D genomic structure and provides a reliable and quantitative understanding of the alterations of genomic structures under different cellular conditions. The genomic structure yielded by CTG serves as an architectural blueprint of the dynamic gene regulatory network, based on which cell-specific correspondence between gene-gene and corresponding protein-protein physical interactions, as well as transcription correlation, is revealed. We also find that gene fusion events are significantly enriched between genes of short CTG distances and are thus close in 3D space. These findings indicate that 3D chromatin structure is at least partially correlated with downstream processes such as transcription, gene regulation, and even regulatory networking through affecting protein-protein interactions.

Clinical Implication of HER2 Aberration in Patients With Metastatic Cancer Using Next-Generation Sequencing: A Pan-Tumor Analysis

PURPOSE

Human epidermal growth factor receptor 2 (HER2) protein expression or gene amplification is a significant predictive biomarker for identifying patients with cancer, who may benefit from HER2-targeted therapy. The aim of this study was to survey the proportion of patients who had HER2 aberration and to investigate the correlation between HER2 amplification and HER2 overexpression in immunohistochemistry (IHC) as a real-world data.

METHODS

We surveyed the incidence of HER2 aberration including mutation (single-nucleotide variant [SNV]), amplification (copy-number variation), and fusion by next-generation sequencing (NGS) in 2,119 patients with cancer from Samsung Medical Center in South Korea.

RESULTS

Of 2,119 patients with cancer, 189 patients (8.9%) had HER2 aberration in their tumor specimen. Of 189 patients, 113 (5.3%) patients had HER2 amplification, 82 (3.9%) patients had HER2 mutations, and 11 (0.5%) patients had HER2 fusion. Of note, 10 patients (0.5%) had concurrent HER2 amplification and HER2 fusion. In addition, we identified that HER2 protein overexpression was strongly related to HER2 amplification by NGS. Of 74 patients with HER2 amplification only by NGS test, 64 patients (86.5%) had HER2 overexpression by IHC. Of 10 patients with concurrent HER2 amplification and fusion, 80% patients were HER2 overexpression. Among 51 patients with only HER2 mutation (SNV), 9 patients (17.6%) were HER2 (+). Interestingly, almost all patients with colorectal cancer (11 of 12) with HER2 amplification had very strong HER2 overexpression (3+) in their tumor specimen.

CONCLUSION

In conclusion, we showed that when patients with metastatic cancer receive NGS test, approximately 8.9% have HER2 aberrations in their tumor specimen. Most patients have HER2 amplification, and a small percentage of patients have HER2 fusion. A great majority of patients with HER2 amplification and/or HER2 fusion had HER2 (+) tumor by IHC.

Characterization and evaluation of gene fusions as a measure of genetic instability and disease prognosis in prostate cancer

BACKGROUND

Prostate cancer (PCa) is one of the most prevalent cancers worldwide. The clinical manifestations and molecular characteristics of PCa are highly variable. Aggressive types require radical treatment, whereas indolent ones may be suitable for active surveillance or organ-preserving focal therapies. Patient stratification by clinical or pathological risk categories still lacks sufficient precision. Incorporating molecular biomarkers, such as transcriptome-wide expression signatures, improves patient stratification but so far excludes chromosomal rearrangements. In this study, we investigated gene fusions in PCa, characterized potential novel candidates, and explored their role as prognostic markers for PCa progression.

METHODS

We analyzed 630 patients in four cohorts with varying traits regarding sequencing protocols, sample conservation, and PCa risk group. The datasets included transcriptome-wide expression and matched clinical follow-up data to detect and characterize gene fusions in PCa. With the fusion calling software Arriba, we computationally predicted gene fusions. Following detection, we annotated the gene fusions using published databases for gene fusions in cancer. To relate the occurrence of gene fusions to Gleason Grading Groups and disease prognosis, we performed survival analyses using the Kaplan-Meier estimator, log-rank test, and Cox regression.

RESULTS

Our analyses identified two potential novel gene fusions, MBTTPS2,L0XNC01::SMS and AMACR::AMACR. These fusions were detected in all four studied cohorts, providing compelling evidence for the validity of these fusions and their relevance in PCa. We also found that the number of gene fusions detected in a patient sample was significantly associated with the time to biochemical recurrence in two of the four cohorts (log-rank test, p-value < 0.05 for both cohorts). This was also confirmed after adjusting the prognostic model for Gleason Grading Groups (Cox regression, p-values < 0.05).

CONCLUSIONS

Our gene fusion characterization workflow revealed two potential novel fusions specific for PCa. We found evidence that the number of gene fusions was associated with the prognosis of PCa. However, as the quantitative correlations were only moderately strong, further validation and assessment of clinical value is required before potential application.

Deep genomic characterization highlights complexities and prognostic markers of pediatric acute myeloid leukemia

Pediatric acute myeloid leukemia (AML) is an uncommon but aggressive hematological malignancy. The poor outcome is attributed to inadequate prognostic classification and limited treatment options. A thorough understanding on the genetic basis of pediatric AML is important for the development of effective approaches to improve outcomes. Here, by comprehensively profiling fusion genes as well as mutations and copy number changes of 141 myeloid-related genes in 147 pediatric AML patients with subsequent variant functional characterization, we unveil complex mutational patterns of biological relevance and disease mechanisms including MYC deregulation. Also, our findings highlight TP53 alterations as strong adverse prognostic markers in pediatric AML and suggest the core spindle checkpoint kinase BUB1B as a selective dependency in this aggressive subgroup. Collectively, our present study provides detailed genomic characterization revealing not only complexities and mechanistic insights into pediatric AML but also significant risk stratification and therapeutic strategies to tackle the disease.

Clinically relevant fusion oncogenes: detection and practical implications

Mechanistically, chimeric genes result from DNA rearrangements and include parts of preexisting normal genes combined at the genomic junction site. Some rearranged genes encode pathological proteins with altered molecular functions. Those which can aberrantly promote carcinogenesis are called fusion oncogenes. Their formation is not a rare event in human cancers, and many of them were documented in numerous study reports and in specific databases. They may have various molecular peculiarities like increased stability of an oncogenic part, self-activation of tyrosine kinase receptor moiety, and altered transcriptional regulation activities. Currently, tens of low molecular mass inhibitors are approved in cancers as the drugs targeting receptor tyrosine kinase (RTK) oncogenic fusion proteins, that is, including ALK, ABL, EGFR, FGFR1-3, NTRK1-3, MET, RET, ROS1 moieties. Therein, the presence of the respective RTK fusion in the cancer genome is the diagnostic biomarker for drug prescription. However, identification of such fusion oncogenes is challenging as the breakpoint may arise in multiple sites within the gene, and the exact fusion partner is generally unknown. There is no gold standard method for RTK fusion detection, and many alternative experimental techniques are employed nowadays to solve this issue. Among them, RNA-seq-based methods offer an advantage of unbiased high-throughput analysis of only transcribed RTK fusion genes, and of simultaneous finding both fusion partners in a single RNA-seq read. Here we focus on current knowledge of biology and clinical aspects of RTK fusion genes, related databases, and laboratory detection methods.

Clinical Implications and Treatment Strategies for ESR1 Fusions in Hormone Receptor-Positive Metastatic Breast Cancer: A Case Series

In hormone receptor-positive metastatic breast cancer (HR+ MBC), endocrine resistance is commonly due to genetic alterations of ESR1, the gene encoding estrogen receptor alpha (ERα). While ESR1 point mutations (ESR1-MUT) cause acquired resistance to aromatase inhibition (AI) through constitutive activation, far less is known about the molecular functions and clinical consequences of ESR1 fusions (ESR1-FUS). This case series discusses 4 patients with HR+ MBC with ESR1-FUS in the context of the existing ESR1-FUS literature. We consider therapeutic strategies and raise the hypothesis that CDK4/6 inhibition (CDK4/6i) may be effective against ESR1-FUS with functional ligand-binding domain swaps. These cases highlight the importance of screening for ESR1-FUS in patients with HR+ MBC while continuing investigation of precision treatments for these genomic rearrangements.

FusionAI, a DNA-sequence-based deep learning protocol reduces the false positives of human fusion gene prediction

Even though there were many tool developments of fusion gene prediction from NGS data, too many false positives are still an issue. Wise use of the genomic features around the fusion gene breakpoints will be helpful to identify reliable fusion genes efficiently. For this aim, we developed FusionAI, a deep learning pipeline predicting human fusion gene breakpoints from DNA sequence. FusionAI is freely available via https://compbio.uth.edu/FusionGDB2/FusionAI.

For complete details on the use and execution of this protocol, please refer to Kim et al. (2021b).

•

FusionAI can predict the fusion breakpoints from the given DNA sequence

•

FusionAI can reduce the false positives of the predicted fusion genes by other tools

•

FusionAI can identify the genomic features related to the genomic breakage

•

FusionAI creates a landscape image of 44 human genomic features around the breakpoints