Abstract. Over the past decade, data science and machine learning has grown from a mysterious art form to a staple tool across a variety of fields in academia, business, and government. In this paper, we introduce the concept of tree-based pipeline optimization for automating one of the most tedious parts of machine learning-pipeline design. We implement a Tree-based Pipeline Optimization Tool (TPOT) and demonstrate its effectiveness on a series of simulated and real-world genetic data sets. In particular, we show that TPOT can build machine learning pipelines that achieve competitive classification accuracy and discover novel pipeline operators-such as synthetic feature constructors-that significantly improve classification accuracy on these data sets. We also highlight the current challenges to pipeline optimization, such as the tendency to produce pipelines that overfit the data, and suggest future research paths to overcome these challenges. As such, this work represents an early step toward fully automating machine learning pipeline design.

Purpose Metastatic breast cancers continue to elude current therapeutic strategies, including those utilizing PI3K inhibitors. Given the prominent role of PI3Kα,β in tumor growth and PI3Kγ,δ in immune cell function, we sought to determine whether PI3K inhibition altered anti-tumor immunity. Experimental Design The effect of PI3K inhibition on tumor growth, metastasis, and anti-tumor immune response was characterized in mouse models utilizing orthotopic implants of 4T1 or PyMT mammary tumors into syngeneic or PI3Kγ null mice, and patient-derived breast cancer xenografts in humanized mice. Tumor infiltrating leukocytes were characterized by IHC and FACS analysis in BKM120 (30mg/kg, QD) or vehicle treated mice and PI3Kγnull versus PI3KγWT mice. Based on the finding that PI3K inhibition resulted in a more inflammatory tumor leukocyte infiltrate, the therapeutic efficacy of BKM120 (30mg/kg, QD) and anti-PD1 (100µg, twice weekly) was evaluated in PyMT tumor bearing mice. Results Our findings show that PI3K activity facilitates tumor growth and surprisingly, restrains tumor immune surveillance. These activities could be partially suppressed by BKM120 or by genetic deletion of PI3Kγ in the host. The anti-tumor effect of PI3Kγ loss in host, but not tumor, was partially reversed by CD8+T cell depletion. Treatment with therapeutic doses of both BKM120 and antibody to PD-1 resulted in consistent inhibition of tumor growth compared to either agent alone. Conclusions PI3K inhibition slows tumor growth, enhances anti-tumor immunity, and heightens susceptibility to immune checkpoint inhibitors. We propose that combining PI3K inhibition with anti-PD1 may be a viable therapeutic approach for triple negative breast cancer.

Recent reports hypothesize that multiple variant DNA repair gene interactions influence cancer susceptibility. However, studies identifying high-risk cancer-related genes use single gene approaches that lack the statistical rigor to model higher-order interactions. To address this issue, we systematically evaluated individual and joint modifying effects of commonly studied polymorphic base and nucleotide excision repair genes relative to prostate cancer (PCA) risk using conventional logistic regression models and multifactor dimensionality reduction (MDR). We hypothesized that inheriting two or more compromised DNA repair loci may increase PCA risk due to altered gene product function. Six genetic alterations were evaluated using germ-line DNA samples from 208 PCA cases and 665 disease-free controls via TaqMan Polymerase Chain Reaction. With the exception of XPD 312, no association existed between individual DNA repair single nucleotide polymorphisms (SNPs) and PCA. Individuals with the XPD 312 Asn/Asn genotype had an 8.6-fold increase in risk (OR = 8.59; 95% CI =1.81–40.66). We did not observe any significant single gene or gene-gene interactions based on MDR modeling. Our findings emphasize the importance of utilizing a combination of traditional and advanced statistical tools to identify and validate single gene and multi-locus interactions in relation to cancer susceptibility.

BackgroundPolymorphisms in glutathione S-transferase (GST) genes may influence response to oxidative stress and modify prostate cancer (PCA) susceptibility. These enzymes generally detoxify endogenous and exogenous agents, but also participate in the activation and inactivation of oxidative metabolites that may contribute to PCA development. Genetic variations within selected GST genes may influence PCA risk following exposure to carcinogen compounds found in cigarette smoke and decreased the ability to detoxify them. Thus, we evaluated the effects of polymorphic GSTs (M1, T1, and P1) alone and combined with cigarette smoking on PCA susceptibility.MethodsIn order to evaluate the effects of GST polymorphisms in relation to PCA risk, we used TaqMan allelic discrimination assays along with a multi-faceted statistical strategy involving conventional and advanced statistical methodologies (e.g., Multifactor Dimensionality Reduction and Interaction Graphs). Genetic profiles collected from 873 men of African-descent (208 cases and 665 controls) were utilized to systematically evaluate the single and joint modifying effects of GSTM1 and GSTT1 gene deletions, GSTP1 105 Val and cigarette smoking on PCA risk.ResultsWe observed a moderately significant association between risk among men possessing at least one variant GSTP1 105 Val allele (OR = 1.56; 95%CI = 0.95-2.58; p = 0.049), which was confirmed by MDR permutation testing (p = 0.001). We did not observe any significant single gene effects among GSTM1 (OR = 1.08; 95%CI = 0.65-1.82; p = 0.718) and GSTT1 (OR = 1.15; 95%CI = 0.66-2.02; p = 0.622) on PCA risk among all subjects. Although the GSTM1-GSTP1 pairwise combination was selected as the best two factor LR and MDR models (p = 0.01), assessment of the hierarchical entropy graph suggested that the observed synergistic effect was primarily driven by the GSTP1 Val marker. Notably, the GSTM1-GSTP1 axis did not provide additional information gain when compared to either loci alone based on a hierarchical entropy algorithm and graph. Smoking status did not significantly modify the relationship between the GST SNPs and PCA.ConclusionA moderately significant association was observed between PCA risk and men possessing at least one variant GSTP1 105 Val allele (p = 0.049) among men of African descent. We also observed a 2.1-fold increase in PCA risk associated with men possessing the GSTP1 (Val/Val) and GSTM1 (*1/*1 + *1/*0) alleles. MDR analysis validated these findings; detecting GSTP1 105 Val (p = 0.001) as the best single factor for predicting PCA risk. Our findings emphasize the importance of utilizing a combination of traditional and advanced statistical tools to identify and validate single gene and multi-locus interactions in relation to cancer susceptibility.