Novel cancer gene candidates significantly overlapped with orthogonal systematic cancer screen hits, supporting the power of this approach to identify cancer genes .

Of the 94 genes encoding at least one SMD, 40 (42%) had already been implicated in cancer according to the Sanger Cancer Gene Census (‘Cancer Census’) [39, 40], including well-established cancer-caus-ing genes such as KRAS, EGFR and TP53.

We compared the resulting novel cancer gene candidates with cancer gene candidates emerging from a large-scale in vivo (mouse) screen Via mutagenesis with Sleeping Beauty trans-posons [45].

Of the subset of 54 candidate genes not already known to be cancer genes , 10 were found in the Sleeping Beauty dataset (~ 2-fold enrichment; P-Value 2 5X 10—3, Fisher’s Exact test, Table 4).

These 52 genes were enriched for evidence of involvement in cancer, with 16 being Cancer Census genes (enrichment factor ~ 11.9; P-value = 6.7 X1043, Fisher’s Exact test), and 15 being candidate cancer genes according to the Sleeping Beauty screen (enrichment factor ~ 4.5; P-value = 1.9 X10'6, Fisher’s Exact test).

Domain-associated mutational biases have been reported in several studies focusing on single well-known cancer genes such as the PI3KCA gene in colon and breast cancer[32], and the NOTCH1 gene in leukemia, breast and ovarian cancer [53].

Here, we analyzed the distribution of somatic missense mutations for 14,083 genes across 21 cancer types and identified 52 genes (36 of which are not yet known to be cancer genes ) for which different domain instances may contribute to different cancer types.

Also, genes containing at one or more SMDs were regarded as candidate cancer genes in this study.

By comparing the domain-level mutational landscapes of different cancers generated by our study to previously reported gene-level mutation landscapes in small cell lung cancer, melanoma, colon cancer, and breast cancer[14, 70—73] , we noticed at least ten cancer-type-spe-cif1c SMDs that do not correspond to any previously reported highly mutated cancer-associat-ed genes.

In addition to correspondence of the discovered SMDs to known cancer-relevant domain families, our set of novel driver gene candidates overlapped significantly with a large-scale screen for cancer genes based on transposon mutagenesis in mouse.

CMapBatch meta-analysis strategy: From individual cancer gene signatures to candidate therapeutics

Twenty-one lung cancer gene signatures (tumour vs. normal comparisons).

For this test, we randomly assigned the 21 lung cancer gene signatures to two groups, one with 10 and the other

We call this set of 23 drugs that transcriptionally reverse lung cancer gene changes and slow growth in lung cancer cell lines—TOP drugs (S2 Table); in subsequent sections, we prioritize significant drugs that have not been tested in NCI-6O by linking them to TOP drugs using a variety of metrics.

There are 4 drugs targeting PLA2G4A included in the CMap collection, and all 4 significantly reverse lung cancer gene changes in our analyses: flunisolide, fluocinonide, fluorometholone, and medrysone.

247 significant drugs consistently reverse lung cancer gene changes in rank prod-82 Table.

The columns COSMIC, TSgene, and ncg denote Whether the gene is present in COSMIC, TSGene, or the Network of Cancer Genes respectively.

The columns COSMIC, TSgene, and ncg denote Whether the gene is present in COSMIC, TSGene, or the Network of Cancer Genes respectively.

The columns COSMIC, TSgene, and ncg denote Whether the gene is present in COSMIC, TSGene, or the Network of Cancer Genes respectively.