DeeprankCDM: A Deep learning Based Rank Aggregation Method for Ranking Mutations in Cancer Driver Gene based on Pathogenicity

Background: Due to a lack of prior technologies able to generate a landscape analysis of mutational neoantigen-HLA class I complex T cell recognition, it is unclear how the frequency of T cell receptors (TCR) against patient-specific mutations compares to shared oncogenic mutations. Methods: We used the imPACT Isolation Technology®, which allows the selective capture of cancer-specific CD8 T cells from the blood of cancer patients at frequencies as low as 1 in 300K CD8 T cells. This TCR isolation platform is based on whole exome sequencing of a tumor and paired normal tissue control for the identification of non-synonymous mutations, including mutations in cancer driver genes (e.g., p53, KRAS, PI3KCA or HRAS). Bioinformatic algorithms were used to predict potential HLA-binding neoantigens (neoAg) in context of the patient’s HLA class I haplotype. A library of multimerized and barcoded predicted neoAg peptide-HLA molecules were generated to interrogate patient CD8 T cells. The paired αβ-TCR sequences were derived from the single cell sorted T cells. TCRs from antigen experienced CD8 T cell were functionally validated by generating T cells expressing neoTCRs, using non-viral precision gene editing to insert the transgenic TCR chains into the endogenous TRAC locus. We previously showed that all (40/40) TCRs isolated using this approach, reacted to patient-matched cancer cell lines (Puig Saus et al. AACR 2020). Results: Neoantigen-specific T cells isolated from the peripheral blood of 243 patients with melanoma, bladder, endometrial, ovarian, colorectal, head and neck, urothelial and breast cancers were analyzed For each patient, an average of 352 neoAg-HLA capture complexes were predicted. Across all patients 58,058 neoAg-HLA complexes were manufactured, spanning 8,804 unique mutations. Among these, only about 2% represented known oncogene driver mutations. The predicted peptide-HLA binding affinity for cancer driver mutations was significantly lower (P<0.001) compared to patient-specific neoAg. Remarkably, functional characterization of 206 TCRs showed that most TCRs (93.7%) recognized patient-specific neoAg. Despite the low percentage of neoAg-HLA capture complexes covering oncogene driver mutations (~ 2%), 6.3% of TCRs (13) targeted known cancer driver proteins. Primary human T cells were then engineered to express each TCR. TCR T cells became functionally activated only when exposed to the cognate antigen, irrespective of whether they target neoAg or cancer driver mutations, confirming the high degree of specificity of the isolated TCRs. Conclusion: A landscape analysis of neoAg-specific TCRs recognizing mutations in cancer has revealed that most T cells recognizing neoantigen-HLA complexes are specific for private mutations in each cancer, though low frequencies of TCR targeting driver mutations can still be identified. Citation Format: Barbara Sennino, Jyoti Mathur, Eva Huang, Andrew Conroy, Marc Ting, Benjamin Yuen, Zheng Pan, Eric Stawiski, Stefanie Mandl. Circulating tumor-specific T cells preferentially recognize patient-specific mutational neoantigens and infrequently recognize shared cancer driver mutations [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 563.

To detect the somatic mutations in complete mitochondrial genome and to investigate the role of mtDNA mutation in the tumorigenesis of lung cancer. DNA were extracted from sixteen lung cancer and corresponding normal tissues. The entire mitochondrial genome was amplified with 32 pairs of overlapping primers. mtDNA mutations were screened by temporal temperature gradient gel electrophoresis. mtDNA fragments showing different banding patterns between tumor and paracancerous tissues were sequenced to identify the exact mutations. The common 4 977 bp deletion was also analyzed in all sixteen tumor tissues as well as the matched paracancerous samples by PCR. Ten out of sixteen (62.5%) tumor tissues showed a total of 29 mutations. Half (5/10) of tumors with somatic mutation presented one mutation and the rests had multiple ones. Out of 29 mutations, 17 (58.62%) were in D-loop region, 2 (6.90%) in rRNA and 10 (34.48%) in mRNA. Among 10 mRNA mutations, 7 were silence and 3 were missense mutations. Five out of twenty-nine alterations were heteroplasmic to heteroplasmic change, one was homoplasmic to heteroplasmic and the remains (23/29, 79.3%) were homoplasmic to homoplasmic change. There were five common deletions found, two in tumor tissues and three in paracancerous tissues. There was no mitochondrial microsatellite instability, except for the short deletion or insertion in np303-309. The high incidence of mtDNA mutations found in patients with lung cancer suggests that mtDNA alterations might play an important role in tumorigenesis of lung cancer. Further studies should be needed to determine the pathological effects of somatic mtDNA mutations in lung cancer.

<b>Abstract ID 24526</b> <b>Poster Board 232</b> The cytochrome P450 (CYP) enzymes are responsible for the metabolism of many drugs. Single nucleotide polymorphisms (SNPs) have large impacts on the activity of these enzymes and result in altered metabolism for a variety of drugs. Cancer patients frequently suffer from both chronic pain and depression, leading to co-administration of selective serotonin reuptake inhibitors (ssRIs) and opioids. CYP2D6 is necessary to convert ssRIs into their inactive metabolites and to activate several prodrug opioids. Ideally, to avoid drug-drug interactions in cancer patients when prescribing ssRI-opioid combinations, careful analysis of relevant SNPs through PCR in individual patients would be employed. However, this approach remains technically and economically inaccessible to many cancer demographics and clinical settings. To address these limitations and provide genomic-based alternatives, we analyzed the frequency of CYP2D6 mutations in sequenced specimens from a varied pool of cancer patients. The primary goal of this study is to use cancer data trends linked to CYP2D6 DNA sequence disruptions to improve informed selection of ssRI-opioid treatments based on specific patient characteristics. We mined the Catalogue of Somatic Mutations in Cancer (COSMIC) from the Sanger Institute to identify CYP2D6 mutations occurring in patients with various cancers types. We also used the CHASM-3.1 and VEST-4 artificial intelligence algorithms from the Cancer-Related Analysis of Variants Toolkit (CRAVAT) to evaluate the functional significance of missense mutations. Results from CHASM reveal the probability of mutations either being drivers or passengers in cancer, while VEST indicates the likelihood of mutations being pathogenic. We found that 188 of 409 patients showed CYP2D6 mutations localized to the coding region. A majority of these (126) were missense mutations, 18 showed high pathogenicity, and no cancer drivers were identified. The three tissue types most frequently affected by CYP2D6 mutations were the large intestine (60), lung (58), and upper aerodigestive tract (49). The mean age at time of sequencing (MATS) for all tissues was 58.2 years. Three tissues showed significant age differences relative to the MATS of all tissue types: the urinary tract (mean=69.5, p=0.0080), hematopoietic and lymphoid tissue (mean=49.1, p=0.0039), and the stomach (mean=66.0, p=0.0027). The primary histology associated with CYP2D6 mutations was carcinoma (327 of 409 patients). Intriguingly, cancers linked to lymphoid neoplasm histology had a significantly different MATS (mean=41.4 years, p=0.0002) relative to all histology types (mean=58.2). The observed missense mutations in the conserved P450 domain of CYP2D6 suggest structural changes that likely affect protein function, thereby precluding effective metabolism of this enzyme’s drug substrates. Therefore, these results indicate that cautious, personalized, and deliberate prescription should be used when co-administering opioids and ssRIs to patients who meet criteria associated with decreased CYP2D6 metabolism in the cancer contexts herein reported.

The accumulation of somatic mutations is a defining hallmark of cancer. The genomes of solid tumors contain thousands of mutations that are present in most or all of the malignant cells in that tumor. In addition to these clonal mutations, subclonal mutations, those occurring only in a fraction of malignant cells, likely contribute to the phenotypic and morphologic heterogeneity of cancer cells within a tumor. The extent of subclonal mutations in cancer, however, has been difficult to quantify, as the high error rate of next-generation sequencing precludes reliable detection of mutations present in fewer than 5% of cells. Here, we apply the highly accurate Duplex Sequencing methodology to quantify the extent of subclonal mutations in colorectal cancer (CRC) and paired normal mucosa. By sequencing known CRC driver and non-driver genes, we find that colorectal cancers without known DNA repair deficits harbor an extensive complement of subclonal mutations, in addition to the large number of clonal mutations previously identified. We show that normal colonic mucosa also accumulates substantial numbers of subclonal mutations in an age-dependent manner. The frequency of tumor-associated subclonal mutations, however, does not correlate with age, and has a spectrum distinct from that seen in normal tissue, indicating different mechanisms of mutation accumulation are operative in normal and tumor tissue. The frequency of tumor cells harboring unique subclonal mutations is so high that it is likely that every possible neutral somatic point mutation is present by the time a tumor is clinically detectable. We propose that these subclonal mutations likely accumulate in a series of punctuated bursts, making it highly likely that resistant subclones will emerge during chemotherapy. Citation Format: Edward John Paul Fox, Michael W. Schmitt, Kate S. Reid-Bayliss, Robert A. Beckman, Lawrence A. Loeb. Extensive subclonal mutations in human colorectal cancers detected by Duplex Sequencing. [abstract]. In: Proceedings of the AACR Special Conference on Colorectal Cancer: From Initiation to Outcomes; 2016 Sep 17-20; Tampa, FL. Philadelphia (PA): AACR; Cancer Res 2017;77(3 Suppl):Abstract nr A08.

Objective: CENPH, centromere protein H, is one of the constitutive kinetochore proteins. High expression of CENPH has been shown in various forms of cancers; however, studies searching the effect of CENPH mutations in cancers are limited. Therefore, the aim of this study is to investigate the potential effects of the missense mutations of CENPH that have been identified in different cancers.&#x0D; Materials and Methods: Missense CENPH mutations, which have been observed in cancers, were downloaded from the COSMIC v89. The effect of missense mutations was predicted by using PredictSNP1.0. The protein structure of the CENPH protein was generated with I-TASSER and missense mutations were visualized on CENPH protein with UCSF Chimera. Structural effects of selected mutations were assessed with HOPE.&#x0D; Results: 34 missense mutations were observed in human cancers. Of the 34 missense mutations 18 mutations were predicted as deleterious and 16 mutations were predicted as neutral with ranging expected accuracies. Predicted missense mutations showed a scattered pattern on 3D CENPH protein. Two of the predicted deleterious missense mutations with higher expected accuracy were further analyzed and assessed according to amino acid properties.&#x0D; Conclusion: This study provided a systematic analysis and evaluation of missense mutations on a CENPH protein that have been observed in different cancers.

Background: The AKT inhibitor capivasertib was recently approved in the 2L+ setting for patients (pts) with hormone receptor-positive metastatic breast cancer with alterations in PIK3CA, AKT1, and/or PTEN. There is a need to further characterize the landscape of PIK3CA, AKT1, and PTEN mutations in breast cancer and their functions, to improve our understanding and therapeutic targeting of PI3K-driven breast cancer. Methods: 51,767 pts with breast cancer who underwent tumor sequencing using FoundationOne®CDx or FoundationOne® were included in this analysis. We assessed the frequency, type, and pathogenicity of genomic alterations found in PIK3CA, PTEN, and AKT1 overall and stratified by receptor subtype, patient ancestry, and sample type. Co-occurrence with other pathogenic variants was explored using a Fisher’s exact test with a FDR correction. We correlated PTEN mutations with functional data from deep mutational scanning. Results: The most frequently altered PI3K-pathway gene was PIK3CA (37.4%, 19,384/51,767), followed by PTEN (13.5%) then AKT1 (5.4%), including both pathogenic and VUS alterations in all genes. Among 1,024 distinct PIK3CA alterations identified, 690 (67.4%) were missense mutations. The most common missense mutations were H1047R (35.6% of PIK3CA-altered cases), E545K (19.7%), and E542K (11.7%). Among 288 AKT1 alterations identified, 205 (71.2%) were missense mutations. The most common missense mutation was E17K (69.7% of AKT1-altered cases) followed by gene amplification (14.0%). The most common alteration among patients bearing PTEN alterations was gene deletion (37.3% of PTEN-altered cases). Among 1,723 distinct PTEN alterations identified, 721 (41.9%) were frameshift mutations, 494 (28.7%) were missense mutations, and 246 (14.3%) were splice site mutations. To determine function, PTEN mutations were correlated with existing PTEN deep mutational scanning datasets; these data will be presented. PIK3CA pathogenic alterations were more prevalent in estrogen receptor (ER)+/HER2- and HER2+ cases (40.3% and 37.6% of cases) versus triple negative breast cancer (TNBC; 20.9%), consistent with previously described data. Among all pathogenic PIK3CA alterations, those outside the 542, 545, and 1047 codons were more common in TNBC (42%) compared to ER+/HER2- or HER2+ (33%, p=0.002 and 33%, p=0.002). AKT1 mutations were more prevalent in ER+/HER2- disease (6.0%) than HER2+ or TNBC (1.7%, p=4.7x10^-20 and 3.0%, p=3.5x10^-5). In contrast, PTEN alterations were more commonly found in TNBC (17.9%) than in ER+/HER2- and HER2+ cases (11.3%, p=1.8x10^-8 and 3.9%, p=2x10^-57). PIK3CA alterations were enriched in metastases compared to local tumors (39.6% vs 33.2%, p=6x10^-41) and were less prevalent in African American pts than other groups (27.1% v 38.6% in Europeans, p=9.6x10^-81), while PTEN and AKT alterations were similar across ancestries. PIK3CA, AKT1, and PTEN pathogenic alterations are all mutually exclusive to each other. Statistically significant co-occurring mutations included PIK3CA with CDH1, MAP3K1, SOX2, TBX3, and PRKCI; PTEN with FAS, TP53, and RB1; and AKT1 with NF1. Conclusions: Here we describe the clinicogenomic landscape of PIK3CA, AKT1, and PTEN alterations in a large breast cancer cohort. Our results corroborate previously reported findings that PIK3CA alterations are present in around 40% of ER+/HER2- breast cancer and concentrated in 3 hotspots. However, one-third of patients have variants beyond E542, E545, and H1047 with less certainty about targetability. The preponderance of rare PIK3CA mutations and co-occurrences between PIK3CA, AKT1, and PTEN with genes outside the PI3K pathway merits functional investigation. Citation Format: Jacqueline Tao, Saumya Sisoudiya, Smruthy Sivakumar, Ethan Sokol, Neil Vasan. Clinicogenomic landscape and function of PIK3CA, AKT1, and PTEN mutations in breast cancer [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2024; 2024 Dec 10-13; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(12 Suppl):Abstract nr P5-02-20.

One of the key challenges of cancer biology is to catalogue and understand the somatic genomic alterations leading to cancer. Although alternative definitions and search methods have been developed to identify cancer driver genes and mutations, analyses of thousands of cancer genomes return a remarkably similar catalogue of around 300 genes that are mutated in at least one cancer type. Yet, many features of these genes and their role in cancer remain unclear, first and foremost when a somatic mutation is truly oncogenic. In this review, we first summarize some of the recent efforts in completing the catalogue of cancer driver genes. Then, we give an overview of different aspects that influence the oncogenicity of somatic mutations in the core cancer driver genes, including their interactions with the germline genome, other cancer driver mutations, the immune system or their potential role in healthy tissues. In coming years, this research holds promise to illuminate how, when and why cancer driver genes and mutations are really drivers, and thereby move personalized cancer medicine and targeted therapies forward.

TP53 Mutations Negatively Impact Survival of Acute Myeloid Leukemia Patients Treated with Standard Doses of Azacitidine

The problem of aggregating individual rankings to create an overall consensus ranking representative of the group is of longstanding interest in group decision making. The problem arises in situations where a group of k Decision Makers (DMs) are asked to rank order n alternatives. The question is how to combine the DMs' rankings into one consensus ranking. Several different approaches have been suggested to aggregate DM responses into a compromise or consensus ranking, however, none is generally recognised as being the best and the similarity of consensus rankings generated by these algorithms is largely unknown. In this paper, we propose a new Weighted-sum ordinal Consensus ranking Method (WCM) with the weights derived from a Sigmoid function. We run Monte Carlo simulation across a range of k and n to compare the similarity of the consensus rankings generated by our method with the best-known method of Borda-Kendall (BAK; Kendall, M. (1962) Rank correlation methods. New York, NY: Hafner) and two other commonly used techniques proposed by Beck, M.P. and Lin, B.W. (1983) 'Some heuristics for the consensus ranking problem', Computers and Operations Research, Vol. 10,

Zinc is vital for the structure and function of ~3000 human proteins and hence plays key physiological roles. Consequently, impaired zinc homeostasis is associated with various human diseases including cancer. Intracellular zinc levels are tightly regulated by two families of zinc transporters: ZIPs and ZnTs; ZIPs import zinc into the cytosol from the extracellular milieu, or from the lumen of organelles into the cytoplasm. In contrast, the vast majority of ZnTs compartmentalize zinc within organelles, whereas the ubiquitously expressed ZnT1 is the sole zinc exporter. Herein, we explored the hypothesis that qualitative and quantitative alterations in ZnT1 activity impair cellular zinc homeostasis in cancer. Towards this end, we first used bioinformatics to analyze inactivating mutations in ZIPs and ZNTs, catalogued in the COSMIC and gnomAD databases, representing tumor specimens and healthy population controls, respectively. ZnT1, ZnT10, ZIP8, and ZIP10 showed extremely high rates of loss of function mutations in cancer as compared to healthy controls. Analysis of the putative functional impact of missense mutations in ZnT1-ZnT10 and ZIP1-ZIP14, using homologous protein alignment and structural predictions, revealed that ZnT1 displays a markedly increased frequency of predicted functionally deleterious mutations in malignant tumors, as compared to a healthy population. Furthermore, examination of ZnT1 expression in 30 cancer types in the TCGA database revealed five tumor types with significant ZnT1 overexpression, which predicted dismal prognosis for cancer patient survival. Novel functional zinc transport assays, which allowed for the indirect measurement of cytosolic zinc levels, established that wild type ZnT1 overexpression results in low intracellular zinc levels. In contrast, overexpression of predicted deleterious ZnT1 missense mutations did not reduce intracellular zinc levels, validating eight missense mutations as loss of function (LoF) mutations. Thus, alterations in ZnT1 expression and LoF mutations in ZnT1 provide a molecular mechanism for impaired zinc homeostasis in cancer formation and/or progression.

&lt;div&gt;Abstract&lt;p&gt;&lt;b&gt;Purpose:&lt;/b&gt; &lt;i&gt;TP53&lt;/i&gt; missense mutations may help to identify prostate cancer with lethal potential. Here, we preanalytically, analytically, and clinically validated a robust IHC assay to detect subclonal and focal &lt;i&gt;TP53&lt;/i&gt; missense mutations in prostate cancer.&lt;/p&gt;&lt;p&gt;&lt;b&gt;Experimental Design:&lt;/b&gt; The p53 IHC assay was performed in a CLIA-accredited laboratory on the Ventana Benchmark immunostaining system. p53 protein nuclear accumulation was defined as any p53 nuclear labeling in &gt;10% of tumor cells. Fifty-four formalin-fixed paraffin embedded (FFPE) cell lines from the NCI-60 panel and 103 FFPE prostate cancer tissues (88 primary adenocarcinomas, 15 metastases) with known &lt;i&gt;TP53&lt;/i&gt; mutation status were studied. DU145 and VCaP xenografts were subjected to varying fixation conditions to investigate the effects of preanalytic variables. Clinical validation was performed in two partially overlapping radical prostatectomy cohorts.&lt;/p&gt;&lt;p&gt;&lt;b&gt;Results:&lt;/b&gt; p53 nuclear accumulation by IHC was 100% sensitive for detection of &lt;i&gt;TP53&lt;/i&gt; missense mutations in the NCI-60 panel (25/25 missense mutations correctly identified). Lack of p53 nuclear accumulation was 86% (25/29) specific for absence of &lt;i&gt;TP53&lt;/i&gt; missense mutation. In FFPE prostate tumors, the positive predictive value of p53 nuclear accumulation for underlying missense mutation was 84% (38/45), whereas the negative predictive value was 97% (56/58). In a cohort of men who experienced biochemical recurrence after RP, the multivariable HR for metastasis among cases with p53 nuclear accumulation compared with those without was 2.55 (95% confidence interval, 1.1–5.91).&lt;/p&gt;&lt;p&gt;&lt;b&gt;Conclusions:&lt;/b&gt; IHC is widely available method to assess for the presence of deleterious and heterogeneous &lt;i&gt;TP53&lt;/i&gt; missense mutations in clinical prostate cancer specimens. &lt;i&gt;Clin Cancer Res; 23(16); 4693–703. ©2017 AACR&lt;/i&gt;.&lt;/p&gt;&lt;/div&gt;

BackgroundBecause driver mutations provide selective advantage to the mutant clone, they tend to occur at a higher frequency in tumor samples compared to selectively neutral (passenger) mutations. However, mutation frequency alone is insufficient to identify cancer genes because mutability is influenced by many gene characteristics, such as size, nucleotide composition, etc. The goal of this study was to identify gene characteristics associated with the frequency of somatic mutations in the gene in tumor samples.ResultsWe used data on somatic mutations detected by genome wide screens from the Catalog of Somatic Mutations in Cancer (COSMIC). Gene size, nucleotide composition, expression level of the gene, relative replication time in the cell cycle, level of evolutionary conservation and other gene characteristics (totaling 11) were used as predictors of the number of somatic mutations. We applied stepwise multiple linear regression to predict the number of mutations per gene. Because missense, nonsense, and frameshift mutations are associated with different sets of gene characteristics, they were modeled separately. Gene characteristics explain 88% of the variation in the number of missense, 40% of nonsense, and 23% of frameshift mutations. Comparisons of the observed and expected numbers of mutations identified genes with a higher than expected number of mutations– positive outliers. Many of these are known driver genes. A number of novel candidate driver genes was also identified.ConclusionsBy comparing the observed and predicted number of mutations in a gene, we have identified known cancer-associated genes as well as 111 novel cancer associated genes. We also showed that adding the number of silent mutations per gene reported by genome/exome wide screens across all cancer type (COSMIC data) as a predictor substantially exceeds predicting accuracy of the most popular cancer gene predicting tool - MutsigCV.

Objectives: It is clinically important to evaluate the performance of a newly developed blood pressure (BP) measurement method under different measurement conditions. This study aims to evaluate the performance of using deep learning-based method to measure BPs and BP change under non-resting conditions.Materials and methods: Forty healthy subjects were studied. Systolic and diastolic BPs (SBPs and DBPs) were measured under four conditions using deep learning and manual auscultatory method. The agreement between BPs determined by the two methods were analysed under different conditions. The performance of using deep learning-based method to measure BP changes was finally evaluated.Results: There were no significant BPs differences between two methods under all measurement conditions (all p > .1). SBP and DBP measured by deep learning method changed significantly in comparison with the resting condition: decreased by 2.3 and 4.2 mmHg with deeper breathing (both p < .05), increased by 3.6 and 6.4 mmHg with talking, and increased by 5.9 and 5.8 mmHg with arm movement (all p < .05). There were no significant differences in BP changes measured by two methods (all p > .4, except for SBP change with deeper breathing).Conclusion: This study demonstrated that the deep learning method could achieve accurate BP measurement under both resting and non-resting conditions.Key messagesAccurate and reliable blood pressure measurement is clinically important. We evaluated the performance of our developed deep learning-based blood pressure measurement method under resting and non-resting measurement conditions.The deep learning-based method could achieve accurate BP measurement under both resting and non-resting measurement conditions.

MutationAssessor (MA) helps researchers evaluate the likely functional impact of somatic and germline mutations in cancer. It provides an evolution-based functional impact score (FIS) to classify mutations based on their likely effect on protein function. FIS scores are based on analysis of patterns of conservation in protein families (conserved residues) and subfamilies (specificity residues). In this new version (r4) we have (1) refined the combinatorial entropy analysis of conservation patterns, (2) recalculated full-length protein multiple sequence alignments covering a larger fraction of human proteins and making use of the explosive growth of protein sequence data, (3) compared predicted functional impact with the pathogenic-benign classification of sequence variants in curated knowledge bases, such as ClinVar, (4) observed the inverse relationship between predicted high functional impact and variant frequency in germline genome sequences and (5) explore the evaluation of switch-of-function mutational effects. Functional impact of ~4 million somatic amino-acid changing mutations across more than 320K human tumor samples are now available in the widely used cBioPortal for Cancer Genomics.

SummaryMutations in the adenomatous polyposis coli (APC) tumor suppressor gene strongly predispose to development of gastro-intestinal tumors. Central to the tumorigenic events in APC mutant cells is the uncontrolled stabilization and transcriptional activation of the protein β-catenin. Many questions remain as to how APC controls β-catenin degradation. Remarkably, the large C-terminal region of APC, which spans over 2000 amino acids and includes critical regions in downregulating β-catenin, is predicted to be natively unfolded. Here we discuss how this uncommonly large disordered region may help to coordinate the multiple cellular functions of APC. Recently, a significant number of germline and somatic missense mutations in the central region of APC were linked to tumorigenesis in the colon as well as extra-intestinal tissues. We classify and localize all currently known missense mutations in the APC structure. The molecular basis by which these mutations interfere with the function of APC remains unresolved. We propose several mechanisms by which cancer-related missense mutations in the large disordered domain of APC may interfere with tumor suppressor activity. Insight in the underlying molecular events will be invaluable in the development of novel strategies to counter dysregulated Wnt signaling by APC mutations in cancer.