Fusion genes are formed as the result
Fusion genes are formed as the result of either structural chromosomal rearrangement including, primarily, translocation, inversion, amplification and deletion, or non-structural aberrations caused by cis- and trans-splicing or transcriptional read-through. Such events are known to play important roles in the initial steps of tumorigenesis . Canonical structural fusions are featured by neoplastic Garcinol and are the demonstrated ideal markers for cancer cell identification and/or targeting (if oncogenic) such as in the case of NTRK fusions . This offers particular benefits for tumors lacking surface markers such as triple negative breast cancer carcinomas which account for 15–20% of all breast cancer cases and still lack effective targeted therapies with acceptable adverse effects .
It was in 1960 that the first evidence of fusion genes in human cancer was pinpointed . An abnormally small chromosome, namely the Philadelphia chromosome, was found in over 95% chronic myelogenous leukemia (CML) patients, where the q-arms of chromosomes 9 and 22 are mutually translocated and carry the BCR-ABL1 fusion gene (Table 1) . It was shown later that the product of BCR-ABL1 is an aberrant ABL1 kinase consistently active in phosphorylating interleukin-3 receptor and, thus, a stimulant for the rewiring of myeloid cells towards CML . Recently, a novel fusion pair ETV6-ABL1 was included in the Catalogue Of Somatic Mutations In Cancer (COSMIC), which is a rare but recurrent fusion (Table 1) that results in constitutive tyrosine kinase activity (similar to BCR-ABL1 fusion) in many hematological malignancies .
Initially associated with hematological malignancies, gene fusions have now been shown to occur in solid tumors  including, e.g., glioblastoma , melanoma , prostate , breast , lung , colorectal , head and neck  cancers. Next-generation sequencing at the transcriptome level (RNA-Seq) has aided in fusion gene discovery and verification as an unbiased instrument [17,18]. Bioinformatic predictions and computational tools such as FusionFinder  further automated such a process. The number of fusion genes has surged from 358 in 2007  to approximately 20,000 in 2017 , with over 90% identified in the last 7 years due to advances in deep sequencing and detection algorithms . According to November 2017 release (v83) of COSMIC, 18,029 fusions are associated with tumors , and the recurrence rates are 21% ± 7.1%, 28% ± 20% and 6% ± 10%, respectively, for hematological disorders, benign and malignant solid tumors .
Mechanisms forming fusion genes
Clinical relevance of fusion genes
Fusion frequency and pattern The frequency of recurrent fusion genes is much lower than other types of somatic mutations  and anti-correlated with that of other driver mutations . For instance, the EML4-ALK driver fusion occurs at a rate of 6% in lung cancer, and those of driver mutations in KRAS and EGFR are 25% and 23%, respectively . According to COSMIC November 2017 release (v83), approximately 83% fusions have the mutation counterparts in at least one of its fusion partners, and only around 2% mutations have associated fusion genes. Also, the number of observed coding mutations and copy number variations are 4,067,689 and 1,271,436, respectively, which are approximately 226 and 71 times that of the observed gene fusions . These suggest that recurrent fusion genes are rare in comparison with other forms of somatic mutations due to, probably, complicated generation mechanisms, and pathological fusion events are likely highly penetrative with prominent transforming potential given their oncogenic redundancy with driver mutations . The fusion pair and pattern vary considerably among cancer types [43,67]. While the majority of fusion genes pair with one single other gene, a few such as MLL, EWS and ETV6 are highly promiscuous . The gene fusion network in acute myelogenous leukemia is densely connected around a few genes such as MLL, and that of ovarian cancer is much more dispersed . The highest fusion rate occurs in bladder cancer and that of the lowest in thyroid carcinoma according to a survey of gene fusions in TCGA . These warrant distinct strategies in the therapeutic design and differential utility of gene fusions across cancers.