Estrogen plays an essential physiologic role in reproduction and a pathologic one in breast cancer. The completion of the human genome has allowed the identification of the expressed regions of protein-coding genes; however, little is known concerning the organization of their cis-regulatory elements. We have mapped the association of the estrogen receptor (ER) with the complete nonrepetitive sequence of human chromosomes 21 and 22 by combining chromatin immunoprecipitation (ChIP) with tiled microarrays. ER binds selectively to a limited number of sites, the majority of which are distant from the transcription start sites of regulated genes. The unbiased sequence interrogation of the genuine chromatin binding sites suggests that direct ER binding requires the presence of Forkhead factor binding in close proximity. Furthermore, knockdown of FoxA1 expression blocks the association of ER with chromatin and estrogen-induced gene expression demonstrating the necessity of FoxA1 in mediating an estrogen response in breast cancer cells.

We systematically generated large-scale data sets to improve genome annotation for the nematode Caenorhabditis elegans, a key model organism. These data sets include transcriptome profiling across a developmental time course, genome-wide identification of transcription factor–binding sites, and maps of chromatin organization. From this, we created more complete and accurate gene models, including alternative splice forms and candidate noncoding RNAs. We constructed hierarchical networks of transcription factor–binding and microRNA interactions and discovered chromosomal locations bound by an unusually large number of transcription factors. Different patterns of chromatin composition and histone modification were revealed between chromosome arms and centers, with similarly prominent differences between autosomes and the X chromosome. Integrating data types, we built statistical models relating chromatin, transcription factor binding, and gene expression. Overall, our analyses ascribed putative functions to most of the conserved genome.

The mammalian Retinoblastoma (RB) family including pRB, p107, and p130 represses E2F target genes through mechanisms that are not fully understood. In D. melanogaster, RB-dependent repression is mediated in part by the multisubunit protein complex Drosophila RBF, E2F, and Myb (dREAM) that contains homologs of the C. elegans synthetic multivulva class B (synMuvB) gene products. Using an integrated approach combining proteomics, genomics, and bioinformatic analyses, we identified a p130 complex termed DP, RB-like, E2F, and MuvB (DREAM) that contains mammalian homologs of synMuvB proteins LIN-9, LIN-37, LIN-52, LIN-54, and LIN-53/RBBP4. DREAM bound to more than 800 human promoters in G0 and was required for repression of E2F target genes. In S phase, MuvB proteins dissociated from p130 and formed a distinct submodule that bound MYB. This work reveals an evolutionarily conserved multisubunit protein complex that contains p130 and E2F4, but not pRB, and mediates the repression of cell cycle-dependent genes in quiescence.

The open source C/C++ program is available at http://www.tongji.edu.cn/∼zhanglab/GFOLD/index.html

The ability to predict the response of a cancer patient to a therapeutic agent is a major goal in modern oncology that should ultimately lead to personalized treatment. Existing approaches to predicting drug sensitivity rely primarily on profiling of cancer cell line panels that have been treated with different drugs and selecting genomic or functional genomic features to regress or classify the drug response. Here, we propose a dual-layer integrated cell line-drug network model, which uses both cell line similarity network (CSN) data and drug similarity network (DSN) data to predict the drug response of a given cell line using a weighted model. Using the Cancer Cell Line Encyclopedia (CCLE) and Cancer Genome Project (CGP) studies as benchmark datasets, our single-layer model with CSN or DSN and only a single parameter achieved a prediction performance comparable to the previously generated elastic net model. When using the dual-layer model integrating both CSN and DSN, our predicted response reached a 0.6 Pearson correlation coefficient with observed responses for most drugs, which is significantly better than the previous results using the elastic net model. We have also applied the dual-layer cell line-drug integrated network model to fill in the missing drug response values in the CGP dataset. Even though the dual-layer integrated cell line-drug network model does not specifically model mutation information, it correctly predicted that BRAF mutant cell lines would be more sensitive than BRAF wild-type cell lines to three MEK1/2 inhibitors tested.

We introduce the TRUST4 open-source algorithm for reconstruction of immune receptor repertoires in αβ/γδ T cells and B cells from RNA-sequencing (RNA-seq) data. Compared with competing methods, TRUST4 supports both FASTQ and BAM format and is faster and more sensitive in assembling longereven full-length-receptor repertoires. TRUST4 can also call repertoire sequences from single-cell RNA-seq (scRNA-seq) data without V(D)J enrichment, and is compatible with both SMART-seq and 5′ 10x Genomics platforms.Both T and B cells can generate diverse receptor (TCR and BCR, respectively) repertoires, through somatic V(D)J recombination, to recognize various external antigens or tumor neoantigens. Following antigen recognition, BCRs also undergo somatic hypermutations (SHMs) to further improve antigen-binding affinity. Repertoire sequencing has been increasingly adopted in infectious disease 1 , allergy 2 , autoimmune 3 , tumor immuology 4 and cancer immunotherapy 5 studies, but it is an expensive assay and consumes valuable tissue samples. Alternatively, RNA-seq data contain expressed TCR and BCR sequences in tissues or peripheral blood mononuclear cells (PBMC). However, because repertoire sequences from V(D)J recombination and SHM are different from the germline, they are often eliminated in the read-mapping step.Previously we developed the TRUST algorithm 6-8 , utilized to de novo assemble immune receptor repertories directly from tissue or blood RNA-seq data. When applied to The Cancer Genome Atlas (TCGA) tumor RNA-seq data, TRUST revealed profound biological insights into the repertoires of tumor-infiltrating T cells 6 and B cells 8 , as well as their associated tumor immunity. Although less sensitive than TCR-seq and BCR-seq, TRUST is able to identify the abundantly expressed and potentially more clonally expanded TCRs/BCRs in the RNA-seq data that are more likely to be involved in antigen binding 9 . Recent years have also seen other computational methods introduced for immune repertoire construction from RNA-seq data, including V'DJer 10 , MiXCR 11 , CATT 12 and ImRep 13 . These methods focus on reconstruction of complementary-determining region 3 (CDR3), with limited ability to assemble full-length V(D)J receptor sequences, although CDR1 and CDR2 on the V sequence still contribute considerably to antigen recognition and binding. For example, five out of six mutations predicted in a recent study to influence antibody affinity in the acidic tumor environment are located in CDR1 and CDR2 (ref. 14 ), and four out of nine positions contributing most to 4A8 antibody binding to the SARS-CoV-2 spike protein are in CDR1 and CDR2 (ref. 15 ). Therefore, algorithms that can infer full-length immune receptor repertoires can facilitate better receptor-antigen interaction modeling.

In classical Hodgkin lymphoma (cHL), the host antitumor immune response is ineffective. Hodgkin Reed-Sternberg (HRS) cells have multifaceted mechanisms to evade the immune system, including 9p24.1 genetic alterations, overexpression of PD-1 ligands, and associated T-cell exhaustion and additional structural bases of aberrant antigen presentation. The clinical success of PD-1 blockade in cHL suggests that the tumor microenvironment (TME) contains reversibly exhausted T effector cells (Teffs). However, durable responses are observed in patients with β2-microglobulin/major histocompatibility complex (MHC) class I loss on HRS cells, raising the possibility of non-CD8 T cell-mediated mechanisms of efficacy of PD-1 blockade. These observations highlight the need for a detailed analysis of the cHL TME. Using a customized time-of-flight mass cytometry panel, we simultaneously assessed cell suspensions from diagnostic cHL biopsies and control reactive lymph node/tonsil (RLNT) samples. Precise phenotyping of immune cell subsets revealed salient differences between cHLs and RLNTs. The TME in cHL is CD4 T-cell rich, with frequent loss of MHC class I expression on HRS cells. In cHLs, we found concomitant expansion of T helper 1 (Th1)-polarized Teffs and regulatory T cells (Tregs). The cHL Th1 Tregs expressed little or no PD-1, whereas the Th1 Teffs were PD-1 The differential PD-1 expression and likely functional Th1-polarized CD4 Tregs and exhausted Teffs may represent complementary mechanisms of immunosuppression in cHL.

Tumor-infiltrating B cells are an important component in the microenvironment with unclear anti-tumor impacts. We enhanced our previous computational algorithm TRUST to extract the B cell immunoglobulin (Ig) hypervariable regions from bulk tumor RNA-seq data. TRUST assembled over 30 million complementarity-determining region 3 (CDR3s) of the B cell heavy chain (IgH) from The Cancer Genome Atlas (TCGA). Widespread B cell clonal expansions and Ig subclass switch events were observed in diverse human cancers. Prevalent somatic copy number alterations in MICA and MICB genes related to antibody-dependent cell mediated cytotoxicity (ADCC) were identified in tumors with elevated B cell activity. IgG3-1 subclass switch interacts with the B cell receptor affinity maturation and defects in the ADCC pathway. Comprehensive pan-cancer analyses of tumor-infiltrating B cell receptor repertoires identified novel tumor immune evasion mechanisms through genetic alterations. The IgH sequences identified here are potentially useful resources for future development of immunotherapies.