Raphael Gottardo

Orchestrating high-throughput genomic analysis with Bioconductor

Bioconductor is an open-source, open-development software project for the analysis and comprehension of high-throughput data in genomics and molecular biology. The project aims to enable interdisciplinary research, collaboration and rapid development of scientific software. Based on the statistical programming language R, Bioconductor comprises 934 interoperable packages contributed by a large, diverse community of scientists. Packages cover a range of bioinformatic and statistical applications. They undergo formal initial review and continuous automated testing. We present an overview for prospective users and contributors.

Model-based analysis of tiling-arrays for ChIP-chip

We propose a fast and powerful analysis algorithm, titled Model-based Analysis of Tiling-arrays (MAT), to reliably detect regions enriched by transcription factor chromatin immunoprecipitation (ChIP) on Affymetrix tiling arrays (ChIP-chip). MAT models the baseline probe behavior by considering probe sequence and copy number on each array. It standardizes the probe value through the probe model, eliminating the need for sample normalization. MAT uses an innovative function to score regions for ChIP enrichment, which allows robust P value and false discovery rate calculations. MAT can detect ChIP regions from a single ChIP sample, multiple ChIP samples, or multiple ChIP samples with controls with increasing accuracy. The single-array ChIP region detection feature minimizes the time and monetary costs for laboratories newly adopting ChIP-chip to test their protocols and antibodies and allows established ChIP-chip laboratories to identify samples with questionable quality that might contaminate their data. MAT is developed in open-source Python and is available at http://chip.dfci.harvard.edu/∼wli/MAT. The general framework presented here can be extended to other oligonucleotide microarrays and tiling array platforms.

Critical assessment of automated flow cytometry data analysis techniques

Traditional methods for flow cytometry (FCM) data processing rely on subjective manual gating. Recently, several groups have developed computational methods for identifying cell populations in multidimensional FCM data. The Flow Cytometry: Critical Assessment of Population Identification Methods (FlowCAP) challenges were established to compare the performance of these methods on two tasks: (i) mammalian cell population identification, to determine whether automated algorithms can reproduce expert manual gating and (ii) sample classification, to determine whether analysis pipelines can identify characteristics that correlate with external variables (such as clinical outcome). This analysis presents the results of the first FlowCAP challenges. Several methods performed well as compared to manual gating or external variables using statistical performance measures, which suggests that automated methods have reached a sufficient level of maturity and accuracy for reliable use in FCM data analysis.

Automated gating of flow cytometry data via robust model‐based clustering

The capability of flow cytometry to offer rapid quantification of multidimensional characteristics for millions of cells has made this technology indispensable for health research, medical diagnosis, and treatment. However, the lack of statistical and bioinformatics tools to parallel recent high‐throughput technological advancements has hindered this technology from reaching its full potential. We propose a flexible statistical model‐based clustering approach for identifying cell populations in flow cytometry data based on t‐mixture models with a Box–Cox transformation. This approach generalizes the popular Gaussian mixture models to account for outliers and allow for nonelliptical clusters. We describe an Expectation‐Maximization (EM) algorithm to simultaneously handle parameter estimation and transformation selection. Using two publicly available datasets, we demonstrate that our proposed methodology provides enough flexibility and robustness to mimic manual gating results performed by an expert researcher. In addition, we present results from a simulation study, which show that this new clustering framework gives better results in terms of robustness to model misspecification and estimation of the number of clusters, compared to the popular mixture models. The proposed clustering methodology is well adapted to automated analysis of flow cytometry data. It tends to give more reproducible results, and helps reduce the significant subjectivity and human time cost encountered in manual gating analysis.

Exosomes in human semen carry a distinctive repertoire of small non-coding RNAs with potential regulatory functions

Semen contains relatively ill-defined regulatory components that likely aid fertilization, but which could also interfere with defense against infection. Each ejaculate contains trillions of exosomes, membrane-enclosed subcellular microvesicles, which have immunosuppressive effects on cells important in the genital mucosa. Exosomes in general are believed to mediate inter-cellular communication, possibly by transferring small RNA molecules. We found that seminal exosome (SE) preparations contain a substantial amount of RNA from 20 to 100 nucleotides (nts) in length. We sequenced 20–40 and 40–100 nt fractions of SE RNA separately from six semen donors. We found various classes of small non-coding RNA, including microRNA (21.7% of the RNA in the 20–40 nt fraction) as well as abundant Y RNAs and tRNAs present in both fractions. Specific RNAs were consistently present in all donors. For example, 10 (of ∼2600 known) microRNAs constituted over 40% of mature microRNA in SE. Additionally, tRNA fragments were strongly enriched for 5’-ends of 18–19 or 30–34 nts in length; such tRNA fragments repress translation. Thus, SE could potentially deliver regulatory signals to the recipient mucosa via transfer of small RNA molecules.

Combining Mixture Components for Clustering

Model-based clustering consists of fitting a mixture model to data and identifying each cluster with one of its components. Multivariate normal distributions are typically used. The number of clusters is usually determined from the data, often using BIC. In practice, however, individual clusters can be poorly fitted by Gaussian distributions, and in that case model-based clustering tends to represent one non-Gaussian cluster by a mixture of two or more Gaussian distributions. If the number of mixture components is interpreted as the number of clusters, this can lead to overestimation of the number of clusters. This is because BIC selects the number of mixture components needed to provide a good approximation to the density, rather than the number of clusters as such. We propose first selecting the total number of Gaussian mixture components, K, using BIC and then combining them hierarchically according to an entropy criterion. This yields a unique soft clustering for each number of clusters less than or equal to K. These clusterings can be compared on substantive grounds, and we also describe an automatic way of selecting the number of clusters via a piecewise linear regression fit to the rescaled entropy plot. We illustrate the method with simulated data and a flow cytometry dataset. Supplemental materials are available on the journal web site and described at the end of the article.

Plasma IgG to Linear Epitopes in the V2 and V3 Regions of HIV-1 gp120 Correlate with a Reduced Risk of Infection in the RV144 Vaccine Efficacy Trial

Neutralizing and non-neutralizing antibodies to linear epitopes on HIV-1 envelope glycoproteins have potential to mediate antiviral effector functions that could be beneficial to vaccine-induced protection. Here, plasma IgG responses were assessed in three HIV-1 gp120 vaccine efficacy trials (RV144, Vax003, Vax004) and in HIV-1-infected individuals by using arrays of overlapping peptides spanning the entire consensus gp160 of all major genetic subtypes and circulating recombinant forms (CRFs) of the virus. In RV144, where 31.2% efficacy against HIV-1 infection was seen, dominant responses targeted the C1, V2, V3 and C5 regions of gp120. An analysis of RV144 case-control samples showed that IgG to V2 CRF01_AE significantly inversely correlated with infection risk (OR= 0.54, p=0.0042), as did the response to other V2 subtypes (OR=0.60-0.63, p=0.016-0.025). The response to V3 CRF01_AE also inversely correlated with infection risk but only in vaccine recipients who had lower levels of other antibodies, especially Env-specific plasma IgA (OR=0.49, p=0.007) and neutralizing antibodies (OR=0.5, p=0.008). Responses to C1 and C5 showed no significant correlation with infection risk. In Vax003 and Vax004, where no significant protection was seen, serum IgG responses targeted the same epitopes as in RV144 with the exception of an additional C1 reactivity in Vax003 and infrequent V2 reactivity in Vax004. In HIV-1 infected subjects, dominant responses targeted the V3 and C5 regions of gp120, as well as the immunodominant domain, heptad repeat 1 (HR-1) and membrane proximal external region (MPER) of gp41. These results highlight the presence of several dominant linear B cell epitopes on the HIV-1 envelope glycoproteins. They also generate the hypothesis that IgG to linear epitopes in the V2 and V3 regions of gp120 are part of a complex interplay of immune responses that contributed to protection in RV144.

MAST: a flexible statistical framework for assessing transcriptional changes and characterizing heterogeneity in single-cell RNA sequencing data

Single-cell transcriptomics reveals gene expression heterogeneity but suffers from stochastic dropout and characteristic bimodal expression distributions in which expression is either strongly non-zero or non-detectable. We propose a two-part, generalized linear model for such bimodal data that parameterizes both of these features. We argue that the cellular detection rate, the fraction of genes expressed in a cell, should be adjusted for as a source of nuisance variation. Our model provides gene set enrichment analysis tailored to single-cell data. It provides insights into how networks of co-expressed genes evolve across an experimental treatment. MAST is available at https://github.com/RGLab/MAST.

The Thai Phase III HIV Type 1 Vaccine Trial (RV144) Regimen Induces Antibodies That Target Conserved Regions Within the V2 Loop of gp120

The Thai Phase III clinical trial (RV144) showed modest efficacy in preventing HIV-1 acquisition. Plasma collected from HIV-1-uninfected trial participants completing all injections with ALVAC-HIV (vCP1521) prime and AIDSVAX B/E boost were tested for antibody responses against HIV-1 gp120 envelope (Env). Peptide microarray analysis from six HIV-1 subtypes and group M consensus showed that vaccination induced antibody responses to the second variable (V2) loop of gp120 of multiple subtypes. We further evaluated V2 responses by ELISA and surface plasmon resonance using cyclic (Cyc) and linear V2 loop peptides. Thirty-one of 32 vaccine recipients tested (97%) had antibody responses against Cyc V2 at 2 weeks postimmunization with a reciprocal geometric mean titer (GMT) of 1100 (range: 200-3200). The frequency of detecting plasma V2 antibodies declined to 19% at 28 weeks post-last injection (GMT: 110, range: 100-200). Antibody responses targeted the mid-region of the V2 loop that contains conserved epitopes and has the amino acid sequence KQKVHALFYKLDIVPI (HXB2 Numbering sequence 169-184). Valine at position 172 was critical for antibody binding. The frequency of V3 responses at 2 weeks postimmunization was modest (18/32, 56%) with a GMT of 185 (range: 100-800). In contrast, naturally infected HIV-1 individuals had a lower frequency of antibody responses to V2 (10/20, 50%; p=0.003) and a higher frequency of responses to V3 (19/20, 95%), with GMTs of 400 (range: 100-3200) and 3570 (range: 200-12,800), respectively. RV144 vaccination induced antibodies that targeted a region of the V2 loop that contains conserved epitopes. Early HIV-1 transmission events involve V2 loop interactions, raising the possibility that anti-V2 antibodies in RV144 may have contributed to viral inhibition.

Analysis of V2 antibody responses induced in vaccinees in the ALVAC/AIDSVAX HIV-1 vaccine efficacy trial.

The RV144 clinical trial of a prime/boost immunizing regimen using recombinant canary pox (ALVAC-HIV) and two gp120 proteins (AIDSVAX B and E) was previously shown to have a 31.2% efficacy rate. Plasma specimens from vaccine and placebo recipients were used in an extensive set of assays to identify correlates of HIV-1 infection risk. Of six primary variables that were studied, only one displayed a significant inverse correlation with risk of infection: the antibody (Ab) response to a fusion protein containing the V1 and V2 regions of gp120 (gp70-V1V2). This finding prompted a thorough examination of the results generated with the complete panel of 13 assays measuring various V2 Abs in the stored plasma used in the initial pilot studies and those used in the subsequent case-control study. The studies revealed that the ALVAC-HIV/AIDSVAX vaccine induced V2-specific Abs that cross-react with multiple HIV-1 subgroups and recognize both conformational and linear epitopes. The conformational epitope was present on gp70-V1V2, while the predominant linear V2 epitope mapped to residues 165–178, immediately N-terminal to the putative α4β7 binding motif in the mid-loop region of V2. Odds ratios (ORs) were calculated to compare the risk of infection with data from 12 V2 assays, and in 11 of these, the ORs were ≤1, reaching statistical significance for two of the variables: Ab responses to gp70-V1V2 and to overlapping V2 linear peptides. It remains to be determined whether anti-V2 Ab responses were directly responsible for the reduced infection rate in RV144 and whether anti-V2 Abs will prove to be important with other candidate HIV vaccines that show efficacy, however, the results support continued dissection of Ab responses to the V2 region which may illuminate mechanisms of protection from HIV-1 infection and may facilitate the development of an effective HIV-1 vaccine.