False Discovery Rate Assessment Research Articles

False positives in trans-eQTL and co-expression analyses arising from RNA-sequencing alignment errors

Sequence similarity among distinct genomic regions can lead to errors in alignment of short reads from next-generation sequencing. While this is well known, the downstream consequences of misalignment have not been fully characterized. We assessed the potential for incorrect alignment of RNA-sequencing reads to cause false positives in both gene expression quantitative trait locus (eQTL) and co-expression analyses. Trans-eQTLs identified from human RNA-sequencing studies appeared to be particularly affected by this phenomenon, even when only uniquely aligned reads are considered. Over 75% of trans-eQTLs using a standard pipeline occurred between regions of sequence similarity and therefore could be due to alignment errors. Further, associations due to mapping errors are likely to misleadingly replicate between studies. To help address this problem, we quantified the potential for "cross-mapping'' to occur between every pair of annotated genes in the human genome. Such cross-mapping data can be used to filter or flag potential false positives in both trans-eQTL and co-expression analyses. Such filtering substantially alters the detection of significant associations and can have an impact on the assessment of false discovery rate, functional enrichment, and replication for RNA-sequencing association studies.

False positives in trans-eQTL and co-expression analyses arising from RNA-sequencing alignment errors.

Sequence similarity among distinct genomic regions can lead to errors in alignment of short reads from next-generation sequencing. While this is well known, the downstream consequences of misalignment have not been fully characterized. We assessed the potential for incorrect alignment of RNA-sequencing reads to cause false positives in both gene expression quantitative trait locus (eQTL) and co-expression analyses. Trans-eQTLs identified from human RNA-sequencing studies appeared to be particularly affected by this phenomenon, even when only uniquely aligned reads are considered. Over 75\% of trans-eQTLs using a standard pipeline occurred between regions of sequence similarity and therefore could be due to alignment errors. Further, associations due to mapping errors are likely to misleadingly replicate between studies. To help address this problem, we quantified the potential for "cross-mapping'' to occur between every pair of annotated genes in the human genome. Such cross-mapping data can be used to filter or flag potential false positives in both trans-eQTL and co-expression analyses. Such filtering substantially alters the detection of significant associations and can have an impact on the assessment of false discovery rate, functional enrichment, and replication for RNA-sequencing association studies.

Identification of Single Amino Acid Substitutions in Proteogenomics.

An important aim of proteogenomics, which combines data of high throughput nucleic acid and protein analysis, is to reliably identify single amino acid substitutions representing a main type of coding genome variants. Exact knowledge of deviations from the consensus genome can be utilized in several biomedical fields, such as studies of expression of mutated proteins in cancer, deciphering heterozygosity mechanisms, identification of neoantigens in anticancer vaccine production, search for RNA editing sites at the level of the proteome, etc. Generation of this new knowledge requires processing of large data arrays from high-resolution mass spectrometry, where information on single-point protein variation is often difficult to extract. Accordingly, a significant problem in proteogenomic analysis is the presence of high levels of false positive results for variant-containing peptides in the produced results. Here we review recently suggested approaches of high quality proteomics data processing that may provide more reliable identification of single amino acid substitutions, especially contrary to residue modifications occurring in vitro and in vivo. Optimized methods for assessment of false discovery rate save instrumental and computational time spent for validation of interesting findings of amino acid polymorphism by orthogonal methods.

Abstract P6-16-03: Intrinsic subtypes and MRI patterns in brain metastasis associated with breast cancer

Abstract Background: Breast cancer is the 2nd most common cancer to metastasize to the brain. The development of brain metastasis (BM) is associated with a lower median survival compared with other locations of metastasis and carries with it high morbidity and reduced quality of life. There are limited treatment options and virtually no approved targeted therapies for this disease. We have previously reported specific pathways enriched in BM compared to primary tumors and now further this study to examine pathways by intrinsic subtype and location of and number of BM. Methods: Archival FFPE material was obtained from BM using pathology records; clinical data, including MRI images, was retrieved from medical records and institutional tumor registry under an IRB protocol. Tumor DNA/RNA was extracted from 2 mm cores macrodissected from the FFPE tissue blocks using the Qiagen AllPrep DNA/RNA FFPE kit. Gene expression profiling was performed using Affymetrix Human Transcriptome Array 2.0 microarray on BM samples for which sufficient RNA was available. Gene-level expression quantification was derived after RMA normalization using the Affymetrix Transcriptome Analysis Console. PAM50 subtypes were assigned by clustering samples using median-subtracted PAM50 gene expression levels. Differential gene expression was estimated using the non-parametric Mann-Whitney test, followed by assessment of false discovery rate using the Banjamini-Hochberg FDR methodology. Pathway enrichment analysis was performed using the NCI Pathway Interaction Database. Results: Gene expression profiling showed the following intrinsic subtype distribution among all BM: luminal A 32% (19/59), luminal B 31% (18/59), HER2 enriched 7% (4/59), and basal subtype 31% (18/59). At time of development of BM 64% (38/59) of patients presented with a single lesion compared to 36% (21/59) of patients who presented with multiple lesions (p=0.12). Thirty-nine percent (7/18) of patients with basal subtypes were observed to present with multiple BM compared to 61% (11/18) non-basal subtypes (p=0.25). In addition, 12% (13/59) of BM were found to be exclusively dural-based lesions. They appeared more frequently in the luminal subtypes [11/13 vs 2/13; p=0.06]. A total of 314 genes were differentially expressed (Wilcox pval < 0.05; FDR < 0.01) between the basal and luminal subtypes. As expected, we found that the FOXA transcription factor network was up-regulated in luminal when compared to the basal subtype, whereas the FOXM1 transcription factor network was up-regulated in the basals. We also found a total of 28 genes to be significantly differentially expressed (Wilcox pval < 0.05; FDR < 0.01) between the dural and non-dural BM. The beta1 integrin and syndecan-1 pathways were significantly enriched, along with angiogenesis and lymphatic endothelium pathways. Key genes in these pathways (COL1A1, COL1A2, COL3A1, CDH11, were found to be at least 2-fold up-regulated in the dural BM compared to non-dural BM. Conclusion: Identifying pathways that are differentially expressed between intrinsic subtypes may help us develop new targeted therapies to provide more treatment options for breast cancer patients with brain metastasis. Citation Format: Nicole Williams, Vinay Varadan, Aditi Vadodkar, Kristy Miskimen, Stephanie Kim, Shaveta Vinayak, Paula Silverman, Jill Barnholtz-Sloan, Andrew Sloan, Cheryl Thompson, Lisa Rogers, Hannah Gilmore, Lyndsay Harris. Intrinsic subtypes and MRI patterns in brain metastasis associated with breast cancer [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P6-16-03.

Exploring multilocus associations of inflammation genes and colorectal cancer risk using hapConstructor

BackgroundIn candidate-gene association studies of single nucleotide polymorphisms (SNPs), multilocus analyses are frequently of high dimensionality when considering haplotypes or haplotype pairs (diplotypes) and differing modes of expression. Often, while candidate genes are selected based on their biological involvement in a given pathway, little is known about the functionality of SNPs to guide association studies. Investigators face the challenge of exploring multiple SNP models to elucidate which variants, independently or in combination, might be associated with a disease of interest. A data mining module, hapConstructor (freely-available in Genie software) performs systematic construction and association testing of multilocus genotype data in a Monte Carlo framework. Our objective was to assess its utility to guide statistical analyses of haplotypes within a candidate region (or combined genotypes across candidate genes) beyond that offered by a standard logistic regression approach.MethodsWe applied the hapConstructor method to a multilocus investigation of candidate genes involved in pro-inflammatory cytokine IL6 production, IKBKB, IL6, and NFKB1 (16 SNPs total) hypothesized to operate together to alter colorectal cancer risk. Data come from two U.S. multicenter studies, one of colon cancer (1,556 cases and 1,956 matched controls) and one of rectal cancer (754 cases and 959 matched controls).ResultsHapConstrcutor enabled us to identify important associations that were further analyzed in logistic regression models to simultaneously adjust for confounders. The most significant finding (nominal P = 0.0004; false discovery rate q = 0.037) was a combined genotype association across IKBKB SNP rs5029748 (1 or 2 variant alleles), IL6 rs1800797 (1 or 2 variant alleles), and NFKB1 rs4648110 (2 variant alleles) which conferred an ~80% decreased risk of colon cancer.ConclusionsStrengths of hapConstructor were: systematic identification of multiple loci within and across genes important in CRC risk; false discovery rate assessment; and efficient guidance of subsequent logistic regression analyses.

Gene expression signatures in non-small cell lung cancer (NSCLC) according to cigarette smoking history

22082 Background: Epidemiological and experimental studies confirm that cigarette smoking is the principal causal agent of lung cancer. It is estimated that about 2–5% of all lung cancer patients (pts) have never smoked. A knowledge on molecular features of lung cancers not related to smoking might lead to better understanding of their carcinogenesis and allow development of new therapeutic targets. In this study we compared gene expression signatures in non-smoking vs smoking NSCLC pts. Methods: Study group included 70 NSCLC pts, of those 25 never-smokers (WHO criteria) and 45 current or past smokers (mean 49.7 pack-years, range 6–130). All pts underwent curative pulmonary resection. There were more women among non-smokers (p=0.003), whereas other major characteristics did not significantly differ between both groups. Qualitative RT-PCR analysis was performed on mRNA derived from snap frozen tumor specimens. Expression of 27 genes: potentially related to smoking, including kinase receptors, sex hormone receptors, transcription factors and genes involved in HPV infection pathways was investigated. 18S rRNA subunit, POLR2A, ESD and YAP1 were used as normalization genes. Gene expressions were compared by parametric permutation test with Benjamini-Hochberg correction for multiple comparisons (False Discovery Rate assessment). Results: In univariate analysis 3 genes were particularly overexpressed in tumors of non-smokers: RRAD (p=0.0002), TGF-beta receptor-2 (TGFBR2; p=0.002) and progesterone receptor (PgR; p=0.007). TGF-beta receptor-3 (TGFBR3; p=0.02), SOX9 (p=0.02) and androgen receptor (AR; p=0.03) were also significantly overexpressed in these tumors. After correction for the impact of sex, histopathology, stage of disease, and multiple comparisons, RRAD and TGFBR2 were independently correlated with smoking history. Conclusions: NSCLC in non-smokers is characterized by a specific gene expression signature. Overexpression of TGFBR2 provides the basis for developing new targeted therapies. No significant financial relationships to disclose.

Simultaneous inference: When should hypothesis testing problems be combined?

Modern statisticians are often presented with hundreds or thousands of hypothesis testing problems to evaluate at the same time, generated from new scientific technologies such as microarrays, medical and satellite imaging devices, or flow cytometry counters. The relevant statistical literature tends to begin with the tacit assumption that a single combined analysis, for instance, a False Discovery Rate assessment, should be applied to the entire set of problems at hand. This can be a dangerous assumption, as the examples in the paper show, leading to overly conservative or overly liberal conclusions within any particular subclass of the cases. A simple Bayesian theory yields a succinct description of the effects of separation or combination on false discovery rate analyses. The theory allows efficient testing within small subclasses, and has applications to “enrichment,” the detection of multi-case effects.