Archives

  • 2022-09
  • 2022-08
  • 2022-07
  • 2022-05
  • 2022-04
  • 2021-03
  • 2020-08
  • 2020-07
  • 2020-03
  • 2019-11
  • 2019-10
  • 2019-09
  • 2019-08
  • 2019-07
  • br samples versus WT samples and the

    2022-08-31


    samples versus WT samples and (2) the samples having only missense mutations versus WT samples. The analysis revealed 21 TSG 3 cancer cases involving 9 TSGs had a Jaccard Index score > 0.05 based on shared DEGs. This result indicated that deleterious missense mutations in these genes and the inactiva-tion mutations led to a shared trans-effect, and the effect was higher than expected by chance. These TSGs included ATRX (LGG), BAP1 (UVM), CASP8 (HNSC), CDKN2A (HNSC), EP300 (BLCA), KEAP1 (LUAD), RUNX1 (LAML), STK11 (LUAD), and TP53 (multiple cancer types). Figure 6 shows three example genes, TP53, ATRX, and KEAP1. Interestingly, even though those deleterious missense mutations had some trans-effect, 
    they did not display a cis-effect. Taken together, these results suggested that the deleterious mutations in these genes also might lead to inactivation of the corresponding TSGs. Therefore, this feature should be considered in searching TSG inactivation events from future cancer genomic data.
    DISCUSSION
    We performed a systematic analysis and characterized the func-tional features of TSG inactivation events in more than 5,000 tumor genomes of 33 cancer types/subtypes. We described the spectrum of genetic alterations leading to TSG inactivation;
    classified a cis-effect and a trans-effect for each TSG event; and investigated the potential impact of TSG events on the transcriptome, in signaling pathways, and in protein interaction networks. Consistent expression patterns of TSGs were observed across cancer types, and TSG group features over-whelmed the similarity of the cancer lineages (Figure 4A). In addi-tion to the well-studied TSGs, such as RB1 and TP53, our results pinpointed TSGs functioning in various processes of epigenomic regulation (e.g., ARID1A, ARID1B, and KDM6A). These included Nicotine remodeling genes, such as ARID1A, SETD2, KDM5C, and KDM6A, and genes involved in transcrip-tion, such as RUNX1 and GATA3. These results have critical implications for identification and interpretation of TSG inactiva-tion events in cancer.
    There are a few limitations in this study. First, some mutation types could not be included in our framework. For example, loss of heterozygosity (LOH) is a common event associated with inactivation of TSGs in many cancer types (Ryland et al., 2015). LOH includes copy-number loss LOH and copy number neutral LOH. In this study, we did not explicitly model LOH events, because we did not have access to the raw data. The model presented could partially integrate LOH events when a somatic inactivation mutation was detected (either a nonsense SNV or a frameshift indel). In such cases, the mutation would be included in the model as either WT_NS_FS mutation type in L1 (it may be considered equivalent to copy-neutral LOH) or Hetlose_NS_FS mutation type in L2 (it may include copy-loss LOH). However, if a mutation was inherited and was not in the somatic mutation calls, we were unable to include it in the model. In addition, if one sample contained multiple NS_FS mu-tations in one TSG, the sample was considered as inactivated only once and would not be assigned a stronger weight, which might underestimate the severity of the inactivation mutations in some cases. Second, the sample-based test or gene-based test likely excluded some candidate TSGs. Some TSGs also function in DNA damage repair (DDR) deficiency, such as BRCA2, ATM, and TP53. Inactivation of these genes may result in increased genomic instability and increased mutation load. The sample-based test matches the mutation load in randomi-zation sets and hence may falsely reject such TSG/DDR genes. This study focused on transcriptomic impact. A more compre-hensive analysis of DDR genes is described in Knijnenburg et al. (2018).
    Third, we only considered genetic events. Other mecha-nisms, such as epigenomic inactivation through methylation, post-transcriptional regulation, could lead to inactivation of TSGs. In addition, some TSGs also function as oncogenes in certain cases, such as TP53 and PTEN. Distinguishing which role the TSG plays and whether a mutation leads to activation or inactivation of the TSG remains a challenge for future work. Finally, the task to distinguish the impacts of TSG inactivation event is complicated as shown in the analysis. We examined the transcriptional footprint in the forms of a cis- and a trans-effect, global versus local impacts, lineage versus TSG similarities, and downstream pathways and networks. Future work should investigate other forms of potential impacts, such as mutation load and epigenomic profile.