A cloud-based training module for efficient de novo transcriptome assembly using Nextflow and Google cloud.

This study describes the development of a resource module that is part of a learning platform named "NIGMS Sandbox for Cloud-based Learning" (https://github.com/NIGMS/NIGMS-Sandbox). The overall genesis of the Sandbox is described in the editorial NIGMS Sandbox at the beginning of this Supplement. This module delivers learning materials on de novo transcriptome assembly using Nextflow in an interactive format that uses appropriate cloud resources for data access and analysis. Cloud computing is a powerful new means by which biomedical researchers can access resources and capacity that were previously either unattainable or prohibitively expensive. To take advantage of these resources, however, the biomedical research community needs new skills and knowledge. We present here a cloud-based training module, developed in conjunction with Google Cloud, Deloitte Consulting, and the NIH STRIDES Program, that uses the biological problem of de novo transcriptome assembly to demonstrate and teach the concepts of computational workflows (using Nextflow) and cost- and resource-efficient use of Cloud services (using Google Cloud Platform). Our work highlights the reduced necessity of on-site computing resources and the accessibility of cloud-based infrastructure for bioinformatics applications.

Mutations in yeast Pcf11, a conserved protein essential for mRNA 3' end processing and transcription termination, elicit the Environmental Stress Response.

The Pcf11 protein is an essential subunit of the large complex that cleaves and polyadenylates eukaryotic mRNA precursor. It has also been functionally linked to gene-looping, termination of RNA Polymerase II (Pol II) transcripts, and mRNA export. We have examined a poorly characterized but conserved domain (amino acids 142-225) of the Saccharomyces cerevisiae  Pcf11 and found that while it is not needed for mRNA 3' end processing or termination downstream of the poly(A) sites of protein-coding genes, its presence improves the interaction with Pol II and the use of transcription terminators near gene promoters. Analysis of genome-wide Pol II occupancy in cells with Pcf11 missing this region, as well as Pcf11 mutated in the Pol II CTD Interacting Domain, indicates that systematic changes in mRNA expression are mediated primarily at the level of transcription. Global expression analysis also shows that a general stress response, involving both activation and suppression of specific gene sets known to be regulated in response to a wide variety of stresses, is induced in the two pcf11 mutants, even though cells are grown in optimal conditions. The mutants also cause an unbalanced expression of cell wall-related genes that does not activate the Cell Wall Integrity pathway but is associated with strong caffeine sensitivity. Based on these findings, we propose that Pcf11 can modulate the expression level of specific functional groups of genes in ways that do not involve its well-characterized role in mRNA 3' end processing.

Loss of Krüppel-like factor 9 deregulates both physiological gene expression and development.

Krüppel-like factor 9 (Klf9) is a ubiquitously expressed transcription factor that is a feedforward regulator of multiple stress-responsive and endocrine signaling pathways. We previously described how loss of Klf9 function affects the transcriptome of zebrafish larvae sampled at a single time point 5 days post-fertilization (dpf). However, klf9 expression oscillates diurnally, and the sampled time point corresponded to its expression nadir. To determine if the transcriptomic effects of the klf9-/- mutation vary with time of day, we performed bulk RNA-seq on 5 dpf zebrafish embryos sampled at three timepoints encompassing the predawn peak and midmorning nadir of klf9 expression. We found that while the major effects of the klf9-/- mutation that we reported previously are robust to time of day, the mutation has additional effects that manifest only at the predawn time point. We used a published single-cell atlas of zebrafish development to associate the effects of the klf9-/- mutation with different cell types and found that the mutation increased mRNA associated with digestive organs (liver, pancreas, and intestine) and decreased mRNA associated with differentiating neurons and blood. Measurements from confocally-imaged larvae suggest that overrepresentation of liver mRNA in klf9-/- mutants is due to development of enlarged livers.

Macrophage differentiation is marked by increased abundance of the mRNA 3' end processing machinery, altered poly(A) site usage, and sensitivity to the level of CstF64.

Regulation of mRNA polyadenylation is important for response to external signals and differentiation in several cell types, and results in mRNA isoforms that vary in the amount of coding sequence or 3' UTR regulatory elements. However, its role in differentiation of monocytes to macrophages has not been investigated. Macrophages are key effectors of the innate immune system that help control infection and promote tissue-repair. However, overactivity of macrophages contributes to pathogenesis of many diseases. In this study, we show that macrophage differentiation is characterized by shortening and lengthening of mRNAs in relevant cellular pathways. The cleavage/polyadenylation (C/P) proteins increase during differentiation, suggesting a possible mechanism for the observed changes in poly(A) site usage. This was surprising since higher C/P protein levels correlate with higher proliferation rates in other systems, but monocytes stop dividing after induction of differentiation. Depletion of CstF64, a C/P protein and known regulator of polyadenylation efficiency, delayed macrophage marker expression, cell cycle exit, attachment, and acquisition of structural complexity, and impeded shortening of mRNAs with functions relevant to macrophage biology. Conversely, CstF64 overexpression increased use of promoter-proximal poly(A) sites and caused the appearance of differentiated phenotypes in the absence of induction. Our findings indicate that regulation of polyadenylation plays an important role in macrophage differentiation.

GLH-1/Vasa represses neuropeptide expression and drives spermiogenesis in the C. elegans germline.

Germ granules harbor processes that maintain germline integrity and germline stem cell capacity. Depleting core germ granule components in C. elegans leads to the reprogramming of germ cells, causing them to express markers of somatic differentiation in day-two adults. Somatic reprogramming is associated with complete sterility at this stage. The resulting germ cell atrophy and other pleiotropic defects complicate our understanding of the initiation of reprogramming and how processes within germ granules safeguard the totipotency and immortal potential of germline stem cells. To better understand the initial events of somatic reprogramming, we examined total mRNA (transcriptome) and polysome-associated mRNA (translatome) changes in a precision full-length deletion of glh-1, which encodes a homolog of the germline-specific Vasa/DDX4 DEAD-box RNA helicase. Fertile animals at a permissive temperature were analyzed as young adults, a stage that precedes by 24 â€‹h the previously determined onset of somatic reporter-gene expression in the germline. Two significant changes are observed at this early stage. First, the majority of neuropeptide-encoding transcripts increase in both the total and polysomal mRNA fractions, suggesting that GLH-1 or its effectors suppress this expression. Second, there is a significant decrease in Major Sperm Protein (MSP)-domain mRNAs when glh-1 is deleted. We find that the presence of GLH-1 helps repress spermatogenic expression during oogenesis, but boosts MSP expression to drive spermiogenesis and sperm motility. These insights define an early role for GLH-1 in repressing somatic reprogramming to maintain germline integrity.

A Genomically and Clinically Annotated Patient-Derived Xenograft Resource for Preclinical Research in Non-Small Cell Lung Cancer.

Patient-derived xenograft (PDX) models are an effective preclinical in vivo platform for testing the efficacy of novel drugs and drug combinations for cancer therapeutics. Here we describe a repository of 79 genomically and clinically annotated lung cancer PDXs available from The Jackson Laboratory that have been extensively characterized for histopathologic features, mutational profiles, gene expression, and copy-number aberrations. Most of the PDXs are models of non-small cell lung cancer (NSCLC), including 37 lung adenocarcinoma (LUAD) and 33 lung squamous cell carcinoma (LUSC) models. Other lung cancer models in the repository include four small cell carcinomas, two large cell neuroendocrine carcinomas, two adenosquamous carcinomas, and one pleomorphic carcinoma. Models with both de novo and acquired resistance to targeted therapies with tyrosine kinase inhibitors are available in the collection. The genomic profiles of the LUAD and LUSC PDX models are consistent with those observed in patient tumors from The Cancer Genome Atlas and previously characterized gene expression-based molecular subtypes. Clinically relevant mutations identified in the original patient tumors were confirmed in engrafted PDX tumors. Treatment studies performed in a subset of the models recapitulated the responses expected on the basis of the observed genomic profiles. These models therefore serve as a valuable preclinical platform for translational cancer research. SIGNIFICANCE: Patient-derived xenografts of lung cancer retain key features observed in the originating patient tumors and show expected responses to treatment with standard-of-care agents, providing experimentally tractable and reproducible models for preclinical investigations.

Differential effects of RASA3 mutations on hematopoiesis are profoundly influenced by genetic background and molecular variant.

Studies of the severely pancytopenic scat mouse model first demonstrated the crucial role of RASA3, a dual RAS and RAP GTPase activating protein (GAP), in hematopoiesis. RASA3 is required for survival in utero; germline deletion is lethal at E12.5-13.5 due to severe hemorrhage. Here, conditional deletion in hematopoietic stem and progenitor cells (HSPCs) using Vav-iCre recapitulates the null phenotype demonstrating that RASA3 is required at the stem and progenitor level to maintain blood vessel development and integrity and effective blood production. In adults, bone marrow blood cell production and spleen stress erythropoiesis are suppressed significantly upon induction of RASA3 deficiency, leading to pancytopenia and death within two weeks. Notably, RASA3 missense mutations in two mouse models, scat (G125V) and hlb381 (H794L), show dramatically different hematopoietic consequences specific to both genetic background and molecular variant. The mutation effect is mediated at least in part by differential effects on RAS and RAP activation. In addition, we show that the role of RASA3 is conserved during human terminal erythropoiesis, highlighting a potential function for the RASA3-RAS axis in disordered erythropoiesis in humans. Finally, global transcriptomic studies in scat suggest potential targets to ameliorate disease progression.

Klf9 is a key feedforward regulator of the transcriptomic response to glucocorticoid receptor activity.

The zebrafish has recently emerged as a model system for investigating the developmental roles of glucocorticoid signaling and the mechanisms underlying glucocorticoid-induced developmental programming. To assess the role of the Glucocorticoid Receptor (GR) in such programming, we used CRISPR-Cas9 to produce a new frameshift mutation, GR369-, which eliminates all potential in-frame initiation codons upstream of the DNA binding domain. Using RNA-seq to ask how this mutation affects the larval transcriptome under both normal conditions and with chronic cortisol treatment, we find that GR mediates most of the effects of the treatment, and paradoxically, that the transcriptome of cortisol-treated larvae is more like that of larvae lacking a GR than that of larvae with a GR, suggesting that the cortisol-treated larvae develop GR resistance. The one transcriptional regulator that was both underexpressed in GR369- larvae and consistently overexpressed in cortisol-treated larvae was klf9. We therefore used CRISPR-Cas9-mediated mutation of klf9 and RNA-seq to assess Klf9-dependent gene expression in both normal and cortisol-treated larvae. Our results indicate that Klf9 contributes significantly to the transcriptomic response to chronic cortisol exposure, mediating the upregulation of proinflammatory genes that we reported previously.

Regulation of the Ysh1 endonuclease of the mRNA cleavage/polyadenylation complex by ubiquitin-mediated degradation.

Mutation of the essential yeast protein Ipa1 has previously been demonstrated to cause defects in pre-mRNA 3' end processing and growth, but the mechanism underlying these defects was not clear. In this study, we show that the ipa1-1 mutation causes a striking depletion of Ysh1, the evolutionarily conserved endonuclease subunit of the 19-subunit mRNA Cleavage/Polyadenylation (C/P) complex, but does not decrease other C/P subunits. YSH1 overexpression rescues both the growth and 3' end processing defects of the ipa1-1 mutant. YSH1 mRNA level is unchanged in ipa1-1 cells, and proteasome inactivation prevents Ysh1 loss and causes accumulation of ubiquitinated Ysh1. Ysh1 ubiquitination is mediated by the Ubc4 ubiquitin-conjugating enzyme and Mpe1, which in addition to its function in C/P, is also a RING ubiquitin ligase. In summary, Ipa1 affects mRNA processing by controlling the availability of the C/P endonuclease and may represent a regulatory mechanism that could be rapidly deployed to facilitate reprogramming of cellular responses.

An expanded mouse testis transcriptome and mass spectrometry defines novel proteins.

The testis transcriptome is exceptionally complex. Despite its complexity, previous testis transcriptome analyses relied on a reductive method for transcript identification, thus underestimating transcriptome complexity. We describe here a more complete testis transcriptome generated by combining Tuxedo, a reductive method, and spliced-RUM, a combinatorial transcript-building approach. Forty-two percent of the expanded testis transcriptome is composed of unannotated RNAs with novel isoforms of known genes and novel genes constituting 78 and 9.8% of the newly discovered transcripts, respectively. Across tissues, novel transcripts were predominantly expressed in the testis with the exception of novel isoforms which were also highly expressed in the adult ovary. Within the testis, novel isoform expression was distributed equally across all cell types while novel genes were predominantly expressed in meiotic and post-meiotic germ cells. The majority of novel isoforms retained their protein-coding potential while most novel genes had low protein-coding potential. However, a subset of novel genes had protein-coding potentials equivalent to known protein-coding genes. Shotgun mass spectrometry of round spermatid total protein identified unique peptides from four novel genes along with seven annotated non-coding RNAs. These analyses demonstrate the testis expresses a wide range of novel transcripts that give rise to novel proteins.

Genomic data analysis workflows for tumors from patient-derived xenografts (PDXs): challenges and guidelines.

BACKGROUND: Patient-derived xenograft (PDX) models are in vivo models of human cancer that have been used for translational cancer research and therapy selection for individual patients. The Jackson Laboratory (JAX) PDX resource comprises 455 models originating from 34 different primary sites (as of 05/08/2019). The models undergo rigorous quality control and are genomically characterized to identify somatic mutations, copy number alterations, and transcriptional profiles. Bioinformatics workflows for analyzing genomic data obtained from human tumors engrafted in a mouse host (i.e., Patient-Derived Xenografts; PDXs) must address challenges such as discriminating between mouse and human sequence reads and accurately identifying somatic mutations and copy number alterations when paired non-tumor DNA from the patient is not available for comparison. RESULTS: We report here data analysis workflows and guidelines that address these challenges and achieve reliable identification of somatic mutations, copy number alterations, and transcriptomic profiles of tumors from PDX models that lack genomic data from paired non-tumor tissue for comparison. Our workflows incorporate commonly used software and public databases but are tailored to address the specific challenges of PDX genomics data analysis through parameter tuning and customized data filters and result in improved accuracy for the detection of somatic alterations in PDX models. We also report a gene expression-based classifier that can identify EBV-transformed tumors. We validated our analytical approaches using data simulations and demonstrated the overall concordance of the genomic properties of xenograft tumors with data from primary human tumors in The Cancer Genome Atlas (TCGA). CONCLUSIONS: The analysis workflows that we have developed to accurately predict somatic profiles of tumors from PDX models that lack normal tissue for comparison enable the identification of the key oncogenic genomic and expression signatures to support model selection and/or biomarker development in therapeutic studies. A reference implementation of our analysis recommendations is available at https://github.com/TheJacksonLaboratory/PDX-Analysis-Workflows .

Reference Trait Analysis Reveals Correlations Between Gene Expression and Quantitative Traits in Disjoint Samples.

Systems genetic analysis of complex traits involves the integrated analysis of genetic, genomic, and disease-related measures. However, these data are often collected separately across multiple study populations, rendering direct correlation of molecular features to complex traits impossible. Recent transcriptome-wide association studies (TWAS) have harnessed gene expression quantitative trait loci (eQTL) to associate unmeasured gene expression with a complex trait in genotyped individuals, but this approach relies primarily on strong eQTL. We propose a simple and powerful alternative strategy for correlating independently obtained sets of complex traits and molecular features. In contrast to TWAS, our approach gains precision by correlating complex traits through a common set of continuous phenotypes instead of genetic predictors, and can identify transcript-trait correlations for which the regulation is not genetic. In our approach, a set of multiple quantitative "reference" traits is measured across all individuals, while measures of the complex trait of interest and transcriptional profiles are obtained in disjoint subsamples. A conventional multivariate statistical method, canonical correlation analysis, is used to relate the reference traits and traits of interest to identify gene expression correlates. We evaluate power and sample size requirements of this methodology, as well as performance relative to other methods, via extensive simulation and analysis of a behavioral genetics experiment in 258 Diversity Outbred mice involving two independent sets of anxiety-related behaviors and hippocampal gene expression. After splitting the data set and hiding one set of anxiety-related traits in half the samples, we identified transcripts correlated with the hidden traits using the other set of anxiety-related traits and exploiting the highest canonical correlation (R = 0.69) between the trait data sets. We demonstrate that this approach outperforms TWAS in identifying associated transcripts. Together, these results demonstrate the validity, reliability, and power of reference trait analysis for identifying relations between complex traits and their molecular substrates.

Factors that influence response classifications in chemotherapy treated patient-derived xenografts (PDX).

In this study, we investigated the impact of initial tumor volume, rate of tumor growth, cohort size, study duration, and data analysis method on chemotherapy treatment response classifications in patient-derived xenografts (PDXs). The analyses were conducted on cisplatin treatment response data for 70 PDX models representing ten cancer types with up to 28-day study duration and cohort sizes of 3-10 tumor-bearing mice. The results demonstrated that a 21-day dosing study using a cohort size of eight was necessary to reliably detect responsive models (i.e., tumor volume ratio of treated animals to control between 0.1 and 0.42)-independent of analysis method. A cohort of three tumor-bearing animals led to a reliable classification of models that were both highly responsive and highly nonresponsive to cisplatin (i.e., tumor volume ratio of treated animals to control animals less than 0.10). In our set of PDXs, we found that tumor growth rate in the control group impacted treatment response classification more than initial tumor volume. We repeated the study design factors using docetaxel treated PDXs with consistent results. Our results highlight the importance of defining endpoints for PDX dosing studies when deciding the size of cohorts to use in dosing studies and illustrate that response classifications for a study do not differ significantly across the commonly used analysis methods that are based on tumor volume changes in treatment versus control groups.

Ipa1 Is an RNA Polymerase II Elongation Factor that Facilitates Termination by Maintaining Levels of the Poly(A) Site Endonuclease Ysh1.

The yeast protein Ipa1 was recently discovered to interact with the Ysh1 endonuclease of the pre-mRNA cleavage and polyadenylation (C/P) machinery, and Ipa1 mutation impairs 3'end processing. We report that Ipa1 globally promotes proper transcription termination and poly(A) site selection, but with variable effects on genes depending upon the specific configurations of polyadenylation signals. Our findings suggest that the role of Ipa1 in termination is mediated through interaction with Ysh1, since Ipa1 mutation leads to decrease in Ysh1 and poor recruitment of the C/P complex to a transcribed gene. The Ipa1 association with transcriptionally active chromatin resembles that of elongation factors, and the mutant shows defective Pol II elongation kinetics in vivo. Ysh1 overexpression in the Ipa1 mutant rescues the termination defect, but not the mutant's sensitivity to 6-azauracil, an indicator of defective elongation. Our findings support a model in which an Ipa1/Ysh1 complex helps coordinate transcription elongation and 3' end processing.

Mutant KLF1 in Adult Anemic Nan Mice Leads to Profound Transcriptome Changes and Disordered Erythropoiesis.

Anemic Nan mice carry a mutation (E339D) in the second zinc finger of erythroid transcription factor KLF1. Nan-KLF1 fails to bind a subset of normal KLF1 targets and ectopically binds a large set of genes not normally engaged by KLF1, resulting in a corrupted fetal liver transcriptome. Here, we performed RNAseq using flow cytometric-sorted spleen erythroid precursors from adult Nan and WT littermates rendered anemic by phlebotomy to identify global transcriptome changes specific to the Nan Klf1 mutation as opposed to anemia generally. Mutant Nan-KLF1 leads to extensive and progressive transcriptome corruption in adult spleen erythroid precursors such that stress erythropoiesis is severely compromised. Terminal erythroid differentiation is defective in the bone marrow as well. Principle component analysis reveals two major patterns of differential gene expression predicting that defects in basic cellular processes including translation, cell cycle, and DNA repair could contribute to disordered erythropoiesis and anemia in Nan. Significant erythroid precursor stage specific changes were identified in some of these processes in Nan. Remarkably, however, despite expression changes in large numbers of associated genes, most basic cellular processes were intact in Nan indicating that developing red cells display significant physiological resiliency and establish new homeostatic set points in vivo.

S100A4 Is a Biomarker and Regulator of Glioma Stem Cells That Is Critical for Mesenchymal Transition in Glioblastoma.

Glioma stem cells (GSC) and epithelial-mesenchymal transition (EMT) are strongly associated with therapy resistance and tumor recurrence, but the underlying mechanisms are incompletely understood. Here, we show that S100A4 is a novel biomarker of GSCs. S100A4+ cells in gliomas are enriched with cancer cells that have tumor-initiating and sphere-forming abilities, with the majority located in perivascular niches where GSCs are found. Selective ablation of S100A4-expressing cells was sufficient to block tumor growth in vitro and in vivo We also identified S100A4 as a critical regulator of GSC self-renewal in mouse and patient-derived glioma tumorspheres. In contrast with previous reports of S100A4 as a reporter of EMT, we discovered that S100A4 is an upstream regulator of the master EMT regulators SNAIL2 and ZEB along with other mesenchymal transition regulators in glioblastoma. Overall, our results establish S100A4 as a central node in a molecular network that controls stemness and EMT in glioblastoma, suggesting S100A4 as a candidate therapeutic target. Cancer Res; 77(19); 5360-73. ©2017 AACR.

Promiscuous DNA-binding of a mutant zinc finger protein corrupts the transcriptome and diminishes cell viability.

The rules of engagement between zinc finger transcription factors and DNA have been partly defined by in vitro DNA-binding and structural studies, but less is known about how these rules apply in vivo. Here, we demonstrate how a missense mutation in the second zinc finger of Krüppel-like factor-1 (KLF1) leads to degenerate DNA-binding specificity in vivo, resulting in ectopic transcription and anemia in the Nan mouse model. We employed ChIP-seq and 4sU-RNA-seq to identify aberrant DNA-binding events genome wide and ectopic transcriptional consequences of this binding. We confirmed novel sequence specificity of the mutant recombinant zinc finger domain by performing biophysical measurements of in vitro DNA-binding affinity. Together, these results shed new light on the mechanisms by which missense mutations in DNA-binding domains of transcription factors can lead to autosomal dominant diseases.

Temporal dynamics of the developing lung transcriptome in three common inbred strains of laboratory mice reveals multiple stages of postnatal alveolar development.

To characterize temporal patterns of transcriptional activity during normal lung development, we generated genome wide gene expression data for 26 pre- and post-natal time points in three common inbred strains of laboratory mice (C57BL/6J, A/J, and C3H/HeJ). Using Principal Component Analysis and least squares regression modeling, we identified both strain-independent and strain-dependent patterns of gene expression. The 4,683 genes contributing to the strain-independent expression patterns were used to define a murine Developing Lung Characteristic Subtranscriptome (mDLCS). Regression modeling of the Principal Components supported the four canonical stages of mammalian embryonic lung development (embryonic, pseudoglandular, canalicular, saccular) defined previously by morphology and histology. For postnatal alveolar development, the regression model was consistent with four stages of alveolarization characterized by episodic transcriptional activity of genes related to pulmonary vascularization. Genes expressed in a strain-dependent manner were enriched for annotations related to neurogenesis, extracellular matrix organization, and Wnt signaling. Finally, a comparison of mouse and human transcriptomics from pre-natal stages of lung development revealed conservation of pathways associated with cell cycle, axon guidance, immune function, and metabolism as well as organism-specific expression of genes associated with extracellular matrix organization and protein modification. The mouse lung development transcriptome data generated for this study serves as a unique reference set to identify genes and pathways essential for normal mammalian lung development and for investigations into the developmental origins of respiratory disease and cancer. The gene expression data are available from the Gene Expression Omnibus (GEO) archive (GSE74243). Temporal expression patterns of mouse genes can be investigated using a study specific web resource (http://lungdevelopment.jax.org).

TauCstF-64 Mediates Correct mRNA Polyadenylation and Splicing of Activator and Repressor Isoforms of the Cyclic AMP-Responsive Element Modulator (CREM) in Mouse Testis.

Spermatogenesis is coordinated by the spatial and temporal expression of many transcriptional and posttranscriptional factors. The cyclic AMP-responsive element modulator (CREM) gene encodes both activator and repressor isoforms that act as transcription factors to regulate spermiogenesis. We found that the testis-expressed paralog of CstF-64, tauCstF-64 (gene symbol Cstf2t), is involved in a polyadenylation site choice switch of Crem mRNA and leads to an overall decrease of the Crem mRNAs that are generated from internal promoters in Cstf2t(-/-) mice. More surprisingly, loss of tauCstF-64 also leads to alternative splicing of Crem exon 4, which contains an important activation domain. Thus, testis-specific CREMtau2 isoform protein levels are reduced in Cstf2t(-/-) mice. Consequently, expression of 15 CREM-regulated genes is decreased in testes of Cstf2t(-/-) mice at 25 days postpartum. These effects might further contribute to the infertility phenotype of these animals. This demonstrates that tauCstF-64 is an important stage-specific regulator of Crem mRNA processing that modulates the spatial and temporal expression of downstream stage-specific genes necessary for the proper development of sperm in mice.