More than half of human genes use alternative cleavage and polyadenylation (ApA) to generate mRNA transcripts that differ in the lengths of their 3′ untranslated regions (UTRs), thus altering the post-transcriptional fate of the message and likely the protein output. The extent of 3′ UTR variation across tissues and the functional role of ApA remain poorly understood. We developed a sequencing method called 3′-seq to quantitatively map the 3′ ends of the transcriptome of diverse human tissues and isogenic transformation systems. We found that cell type-specific gene expression is accomplished by two complementary programs. Tissue-restricted genes tend to have single 3′ UTRs, whereas a majority of ubiquitously transcribed genes generate multiple 3′ UTRs. During transformation and differentiation, single-UTR genes change their mRNA abundance levels, while multi-UTR genes mostly change 3′ UTR isoform ratios to achieve tissue specificity. However, both regulation programs target genes that function in the same pathways and processes that characterize the new cell type. Instead of finding global shifts in 3′ UTR length during transformation and differentiation, we identify tissue-specific groups of multi-UTR genes that change their 3′ UTR ratios; these changes in 3′ UTR length are largely independent from changes in mRNA abundance. Finally, tissue-specific usage of ApA sites appears to be a mechanism for changing the landscape targetable by ubiquitously expressed microRNAs.

IMPORTANCEThe combination of erlotinib and bevacizumab as initial treatment of epidermal growth factor receptor (EGFR [OMIM 131550])-mutant lung cancers improves progression-free survival (PFS) compared with erlotinib alone. Because osimertinib prolongs PFS compared with erlotinib, this trial was designed to study the combination of osimertinib and bevacizumab as first-line treatment.OBJECTIVES To determine the safety and tolerability of osimertinib and bevacizumab combination treatment and assess the 12-month PFS of the combination in patients with metastatic EGFR-mutant lung cancers. From August 15, 2016, to May 15, 2018 patients with metastatic EGFR-mutant lung cancers were enrolled in this interventional clinical trial, conducted at a single academic cancer center. In the phase 1 portion of the study, a standard 3 + 3 dose de-escalation design was used to determine the maximum tolerated dose of osimertinib and bevacizumab. In the phase 2 portion of the study, patients were treated at the maximum tolerated dose defined in the phase 1 portion. Statistical analysis was performed from August 1 to October 1, 2019. DESIGN, SETTING, AND PARTICIANTSINTERVENTIONS All patients received osimertinib, 80 mg daily, and bevacizumab, 15 mg/kg once every 3 weeks. MAIN OUTCOMES AND MEASURESThe primary objective of the phase 2 portion of the study was to determine the number of patients receiving the combination of osimertinib and bevacizumab who were progression free at 12 months. Secondary end points included overall response rate, median PFS, overall survival, and definition of the toxic effects of the combination treatment. RESULTS Among the 49 patients in the study (34 women; median age, 60 years [range, 36-83 years]), PFS at 12 months was 76% (95% CI, 65%-90%). The overall response rate was 80% (95% CI, 67%-91%), and median PFS was 19 months (95% CI, 15-24 months). Of the 6 patients with measurable central nervous system disease, all had a partial or complete central nervous system response. Persistent detection of EGFR-mutant circulating tumor (ct)DNA at 6 weeks was associated with shorter median PFS (clearance at 6 weeks, 16.2 months [95% CI, 13 months to not reached]; and no clearance at 6 weeks, 9.8 months [95% CI, 4 months to not reached]; P = .04) and median overall survival (clearance at 6 weeks, not reached; and no clearance at 6 weeks, 10.1 months [95% CI, 6 months to not reached]; P = .002). Identified mechanisms of resistance included squamous cell transformation (n = 2) pleomorphic transformation (n = 1), and acquired EGFR L718Q (n = 1) and C797S (n = 1) mutations. CONCLUSIONS AND RELEVANCEThe combination of osimertinib and bevacizumab met the study's prespecified effectiveness end point. Persistent EGFR-mutant circulating tumor DNA at 6 weeks was associated with early progression and shorter survival. A randomized phase 3 study comparing osimertinib and bevacizumab with osimertinib alone is planned.

Predicting the affinity profiles of nucleic acid-binding proteins directly from the protein sequence is a major unsolved problem. We present a statistical approach for learning the recognition code of a family of transcription factors (TFs) or RNA-binding proteins (RBPs) from high-throughput binding assays. Our method, called affinity regression, trains on protein binding microarray (PBM) or RNA compete experiments to learn an interaction model between proteins and nucleic acids, using only protein domain and probe sequences as inputs. By training on mouse homeodomain PBM profiles, our model correctly identifies residues that confer DNA-binding specificity and accurately predicts binding motifs for an independent set of divergent homeodomains. Similarly, learning from RNA compete profiles for diverse RBPs, our model can predict the binding affinities of held-out proteins and identify key RNA-binding residues. More broadly, we envision applying our method to model and predict biological interactions in any setting where there is a high-throughput ‘affinity’ readout.

A high-resolution physical and transcription map has been generated of a 3.5-Mb region of 5p15.2 that is associated with the Cri du chat (CDC) syndrome. Utilizing a variety of resources including a natural deletion panel, a chromosome specific radiation hybrid panel, and YAC, PAC, and BAC genomic clones we have ordered >60 STSs within this region. Approximately 45% of these STSs were obtained from publicly available whole genome maps, thus allowing for integration of this map with currently available resources. Thirteen of these markers were ESTs. In addition, >70 exon trapped products have been mapped on the natural deletion panel and bacterial clone resource. The combination of these resources has allowed for the identification of 17 transcripts within this region, all of which represent candidate genes for CDC. Further characterization of the genomic contig also revealed that this region of 5p15 contains a large number of repetitive elements.[The sequence data described in this paper have been submitted to GenBank under accession nos. G31374-G31412, B07604-B07657. On-line supplementary material concerning markers used, primers, PCR product sizes, and annealing conditions is available at http://www.cshl.org/gr.]Cri du chat (CDC) syndrome is associated with deletions of 5p15 and is one of the most common contiguous gene disorders with an incidence of 1 in 50,000 live births (Niebuhr 1978). Hallmarks of this syndrome include severe mental retardation, speech delay, prenatal and postnatal growth delay, hypotonia, microcephaly, a round face with downslanting palpebral fissures, hypertelorism, epicanthal folds, low-set and/or poorly formed pinnae, broad nasal bridge with prominent nasal root, microretrognathia, and a plaintive, high-pitched cry similar to the mewing of a cat (Lejeune et al. 1963;Niebuhr 1978;Baccichetti et al. 1988;Church et al. 1995). There have been no reports of CDC without a cytogenetically visible chromosome rearrangement, suggesting that several megabases of DNA must be deleted to produce the classical CDC phenotype. Previous cytogenetic studies indicate that there is a CDC critical region in 5p15. 2-15.3 (Niebuhr 1978). A few patients have been described who present with only a subset of characteristics (Baccichetti et al. 1988;Smith et al. 1990;Overhauser et al. 1994;Church et al. 1995). Analyses of the abnormal chromosomes from these individuals have resulted in the correlation of individual phenotypic characteristics with deletions of specific portions of 5p (Overhauser et al. 1994;Church et al. 1995).Recently, attempts have been made to identify the smallest region of DNA commonly deleted in individuals with classical CDC (Overhauser et al. 1994;Church et al. 1995). Reports for the localization of speech delay and the catlike cry have been consistent, with speech delay localizing to the distal part of the terminal cytogenetic band 5p15.3 and the catlike cry localizing to the proximal portion of this cytogenetic band (Baccichetti et al. 1988;Overhauser et al. 1994;Church et al. 1995). H...