Information

Is it possible to establish a newer animal model completely based on Bioinformatics studies?

Is it possible to establish a newer animal model completely based on Bioinformatics studies?


We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

I wanted to know, if it is possible to prove a new organism as an animal model for any human disease by only Bioinformatics methods like NGS and Structural bioinformatics.

For Example, let's say Drosophila is an animal model for Disease X as it have some 60-70% similarity in the gene and protein structure. Can we prove, a newer organism is a better model for X, if it shows a higher similarity, let's say 80-85%, to gene and protein structure of human?

UPDATE (First I posted it as comment! But I think, this is suited here best!)

@David: Thanks for your answer! Actually, I did looked into it, Large animal models of rare genetic disorders: sheep as phenotypically relevant models of human genetic disease. I am more interested in listing down the problems if we do start this kind of project in our lab. Do NGS is a solution or what part of biology we need to focus so, that we can be more accurate.


It is naïve to think that the extent of protein similarity is sufficient to determine what is the best animal model for a human disease. The physiology of the animal and the question of compensatory genes are all factors that contribute.

Indeed, if the protein is functionally similar, it may be irrelevant if it has diverged in other regions. And, of course, there are ethical and cost considerations - the animals closest to humans are the apes!

I suggest you search for literature on the subject. It's not my field, but Googling brought up this paper discussing models for kidney disease, and no doubt you could find much more yourself.


Objective

Consumption of acidic food and drinks is considered as important risk factor for development of dental erosion. There are several in vitro and in situ studies focusing on the risk indicators and preventive treatment, however, the need for a standardized animal model has been emphasised for many years. The aim was to establish an animal model of extrinsic dental erosion, which may serve as a standard for future studies to improve our understanding of the erosion.

Design

Two acidic drinks, sports drink and cola drink, were given to young mice for six weeks. Experimental and control (water) molars and incisors were dissected out and observed by scanning electron microscopy (SEM). Mandibular first molars were subsequently ground transversely and observed again by SEM. The tooth height and enamel thickness were measured on the SEM images.

Results

The lingual surface of the mandibular molars was most eroded after consumption of acidic drinks. The cola drink exhibited higher erosive effect on mandibular molars compared to sports drink. The lingual tooth height, compared to control, was about 34% and 18% lower in the cola drink and sports drink molars, respectively. Compared to the control molars, the lingual enamel was about 23% thinner in the sports drink molars and totally eroded on the certain lingual areas of the cola drink molars.

Conclusions

This new animal model of extrinsic dental erosion and the presented method with ground molars observed in SEM are suitable for further studies, which will gain deeper insights into the erosive disease.


Choosing The Right Animal Model for Renal Cancer Research ☆

The increase in the life expectancy of patients with renal cell carcinoma (RCC) in the last decade is due to changes that have occurred in the area of preclinical studies. Understanding cancer pathophysiology and the emergence of new therapeutic options, including immunotherapy, would not be possible without proper research. Before new approaches to disease treatment are developed and introduced into clinical practice they must be preceded by preclinical tests, in which animal studies play a significant role. This review describes the progress in animal model development in kidney cancer research starting from the oldest syngeneic or chemically-induced models, through genetically modified mice, finally to xenograft, especially patient-derived, avatar and humanized mouse models. As there are a number of subtypes of RCC, our aim is to help to choose the right animal model for a particular kidney cancer subtype. The data on genetic backgrounds, biochemical parameters, histology, different stages of carcinogenesis and metastasis in various animal models of RCC as well as their translational relevance are summarized. Moreover, we shed some light on imaging methods, which can help define tumor microstructure, assist in the analysis of its metabolic changes and track metastasis development.


2 METHODS

2.1 Data preparation

A complete description of the experimental design, data set generation and processing can be found in ( Poussin et al., 2014). In brief, 19 phosphoproteins, 22 cytokines and genome-wide mRNA levels were measured under 52 different stimuli or Dulbecco’s Modified Eagle’s Medium (DME) control treatments (in triplicate), Table 1. The experiment was performed in two parts: 40 stimuli in the first experiment and 12 in the second. In each part, primary NHBE and NRBE cells were grown and exposed to the indicated number of stimuli. Cells were collected and lysed at different time points: 5 and 25 min. For phosphoprotein measurements, 6 h for gene expression measurements and 24 h for cytokine measurements. All cells were exposed to stimuli in triplicate, and DME controls were performed in 4-, 5- or 6-plicate.

Dataset . Condition . Number of replicates . Number of measurements . Time point(s) . Total size .
Phospho-proteomics 52 stimuli 3 biological replicates 18 phosphoproteins 5 min 10 000+ data points
25 min
mRNA expression 20 000 human genes 6 h 330+ CEL files
19 000 rat genes
Cytokine level 22 cytokines 24 h 7000+ data points
Dataset . Condition . Number of replicates . Number of measurements . Time point(s) . Total size .
Phospho-proteomics 52 stimuli 3 biological replicates 18 phosphoproteins 5 min 10 000+ data points
25 min
mRNA expression 20 000 human genes 6 h 330+ CEL files
19 000 rat genes
Cytokine level 22 cytokines 24 h 7000+ data points
Dataset . Condition . Number of replicates . Number of measurements . Time point(s) . Total size .
Phospho-proteomics 52 stimuli 3 biological replicates 18 phosphoproteins 5 min 10 000+ data points
25 min
mRNA expression 20 000 human genes 6 h 330+ CEL files
19 000 rat genes
Cytokine level 22 cytokines 24 h 7000+ data points
Dataset . Condition . Number of replicates . Number of measurements . Time point(s) . Total size .
Phospho-proteomics 52 stimuli 3 biological replicates 18 phosphoproteins 5 min 10 000+ data points
25 min
mRNA expression 20 000 human genes 6 h 330+ CEL files
19 000 rat genes
Cytokine level 22 cytokines 24 h 7000+ data points

mRNA samples from the first experiment were processed in three batches. Each batch included human and rat mRNA for a subset of stimuli. DME control mRNA samples (four replicates) were measured for each batch. For the second experiment, all mRNA samples were processed together, including DME control mRNA samples (five replicates). Low-quality chips were excluded following quality control (QC) analysis. All remaining expression data including two to three replicates per stimulus were normalized using GC robust multiarray averaging within species. Probesets were mapped to gene symbols using Affymetrix annotations: HG-U133 Plus 2 (na33) and Rodent 230 2.0 (na32), for human and rat, respectively. Probesets mapping to multiple genes were excluded. In cases of multiple probesets mapping to the same gene, the probeset with the highest average expression over all experimental conditions was selected as representative. These high-quality normalized gene expression data in the gene symbol namespace were provided to the participants.

Protein phosphorylation status was measured independently for each experiment part in cell lysates collected at 5 and 25 min (in triplicates) using Luminex xMap ( Dunbar, 2006). Experiment parts 1 and 2 have 6 and 5 DME controls, respectively. After QC, 16 phosphoproteins were kept for the challenge. Data were normalized using a robust regression, and normalized values were provided as the ratio of residuals to the root mean squared error of the fit. Cytokine data were similarly processed, though normalization was carried out by taking the Z-score of each cytokine across all stimuli within an experimental batch, including DME controls.

All data were divided into two equal groups, subsets A and B, by stimulus treatment to be used for training and testing of methods. To ensure similar distributions of signals in both data subsets, stimuli were separated through a data-driven approach that clustered stimuli according to phosphorylation level, gene set activation, gene expression (GEx) batch and differential gene expression. For each cluster, stimuli were randomly assigned to subset A or B.

Orthologs were identified using the HGNC Comparison of Orthology Predictions (downloaded December 19, 2012). Only gene symbol mappings between human and rat were used. A total of 12 458 orthologs were common between human and rat Affymetrix arrays after mapping of probesets to gene symbols.

Gene sets were based on the C2CP (Canonical Pathways) collection from MSigDB v3.1 of the Broad Institute ( Subramanian et al., 2005). This collection was filtered to remove highly redundant gene sets, i.e. overlapping gene sets with many shared members, ensuring that remaining gene sets cover as many pathways/biological functions as possible. The resulting 246 gene sets were used for the STC. Gene set enrichment analysis (GSEA) was performed to assess co-regulation of genes representative of pathways/biological functions. For the analysis, genes were ranked based on calculated LIMMA t-values comparing respective DME control versus stimulus conditions ( Smyth, 2004). LIMMA was performed using the lmFit and eBayes functions from the limma R package for the R Statistical Language with default parameters. The design matrix was constructed to compare the batch-specific DME control with each stimulus individually. Computed NES and associated significance values for each gene set were indicative of the activation/perturbation (increase or decrease) of pathways/biological functions by each stimulus in NHBE and NRBE cells ( Subramanian et al., 2005). GSEA size parameters were min = 15 and max = 500. GSEA NES and FDR q-values were provided to participants.

2.2 Scoring

Sub-challenges 1 (SC1), 2 (SC2) and 3 (SC3) were scored as binary classification problems. Starting with the postulate that no single metric will capture all the attributes of a prediction, we used an aggregate of three metrics for evaluation. The metrics were proposed by IBM team members, and an independent panel of experts comprising the External Scoring Panel (ESP) decided on the final scoring approach. Participant identities were kept anonymous to the IBM team scoring the submissions. Five other metrics were considered but rejected as being redundant to the chosen three. The details of these metrics were not disclosed to the participants until the end of the challenge to avoid influencing method development toward optimizing for the scoring function rather than solving the scientific question posed. This practice is in keeping of other prediction evaluation challenges, like CASP, DREAM and a previous iteration of sbv IMPROVER.

We used non-redundant metrics that highlight three different qualities of a prediction: threshold versus non-threshold, order-based versus confidence-based and different ways of rewarding correct versus incorrect predictions. The chosen metrics were also selected to avoid rewarding pathological predictions, e.g. predicting all items to be of one class. Further complicating metric selection, the quantities of both classes (active and inactive) were imbalanced in the STC with active cases accounting for only ∼ 10% of all cases.

Participants were required to give confidence values for their predictions of either protein phosphorylation status or gene set activation (increase or decrease) to a given stimulus, depending on the sub-challenge. Confidence values could range between 0 and 1, where 1 represents the full confidence of an element being activated (either up- or downregulated) and 0 for full confidence of inactivation. A binarized gold standard (GS) was developed for protein phosphorylation status and gene set activation. For the phosphoprotein GS, normalized expression levels, which is akin to the standard deviation of a normal distribution, with an absolute value ≥3 were considered active, as agreed on by the ESP. Similarly for gene set activation, GSEA FDR q-values ≤0.25 were designated active, as recommended by GSEA.

The submitted matrix of predictions (stimuli versus protein or gene set response) could have been scored column-by-column or row-by-row and then aggregated together. However, given the sparseness of the GS for both protein phosphorylation status and gene set activation, we decided (in agreement with the ESP) to transform the matrix into a vector for scoring, i.e. columns of the matrix were joined to obtain single vector.

2.2.1 Metric descriptions

Area Under the PrecisionRecall Curve (AUPR) is a well-known measure of classifier power. A list of items is ordered by descending confidence value (used only for ranking and not directly in the metric). The list is traversed corresponding to increasingly permissive confidence thresholds, and precision (fraction of ‘active’ predictions that are correct) is plotted versus recall (fraction of true ‘active’ class members correctly predicted). The area under this precision–recall curve is the AUPR score and is represented by a single number that summarizes the tradeoff between both measures.

where TP is the number of true positives, P is total number of positives, TN is the number of true negatives and N is the total number of negatives. For the STC, we used a confidence threshold of 0.5 to binarize the predictions as either positive (≥0.5) or negative (<0.5).

For simplicity, we will refer to PCCnormalized as PCC when in reference to the challenge scoring metric.

2.2.2 Metrics aggregation

A rank-sum scheme to aggregate scoring metrics was proposed by the IBM team, along with one alternative, and was selected by the ESP because it equally weights each metric to produce an overall ranking. This rank-sum scheme was composed of ranking all teams within each respective metric. A team’s aggregate rank was then calculated by summing their rank across these three metrics. This rank sum was used for the final ordering of participants, with best performers achieving the lowest rank sum. To determine the robustness of these rankings, bootstrapping was performed to ensure that best performers were not highly sensitive to the exact configuration of GS. GS was sampled without replacement 1000 times, and the rankings recomputed each time. Given the imbalanced nature of GS, the bootstrapping was constrained to maintain the same proportion of active versus inactive items as observed in the entire GS.

2.3 Statistical significance of metrics

The null distribution for each metric in SC1-3 was generated by scoring 10 6 random predictions. To generate the confidences of a random submission, a uniform random number r (0r1) was generated for each ‘item’.

FDRs were computed for each metric using the R ( Computing, 2013) function p.adjust with the method = ‘fdr’, which computes the Benjamini and Hochberg (1995) correction.

To compute a score’s P-value for each of the metrics, we counted the number of random predictions that were better than or equal to the observed score and divided it by the number of simulated predictions. FDR correction ( Benjamini and Hochberg, 1995) was applied to the P-value, and a value of ≤0.05 was considered to be statistically significant.

The measure S represents the overall response similarity between human H and rat R GS, and is a Matthews correlation coefficient (MCC). The MCC represents a Pearson correlation between two binary vectors. A high S value would indicate a putatively conserved response and a signal that is expected to be translatable. Similarity measures can also be calculated per stimulus Ss = MCC(Rs, Hs), where Rs and Hs are binary vectors of phosphoprotein or gene set responses to stimulus s per phosphoprotein Sp = MCC(Rp, Hp), where Rp and Hp are binary vectors of responses to stimuli for phosphoprotein p and per gene set Sg = MCC(Rg, Hg), where Rg and Hg are binary vectors of responses to stimuli for gene set g.

Predictability Pr represents the overall similarity or agreement between the GS and a team’s or aggregate of teams’ predictions T, and is a MCC. A high Pr value would indicate good prediction performance and that the response was predictable. Like S, Pr can be calculated per stimulus Prs = MCC(GSs, Ts), where GSs and Ts are binary vectors of predicted phosphoprotein or gene set responses to stimulus s per phosphoprotein Prp = MCC(GSp, Tp), where GSp and Tp are binary vectors of predicted responses to stimuli for phosphoprotein p and per gene set Prg = MCC(GSg, Tg), where GSg and Tg are binary vectors of predicted responses to stimuli for gene set g.

The empirical P-values for the presence of genes in overlapping gene sets were calculated by sampling 10 5 times choosing a group of 25 gene sets of 246 gene sets. The frequency a gene is a member of the 25 randomly selected gene sets is recorded. The P-value is obtained by dividing the frequency a gene was found in at least x gene sets by 10 5 .


Histological aspects of the “fixed-particle” model of stone formation: animal studies

Crystallization by itself is not harmful as long as the crystals are not retained in the kidneys and are allowed to pass freely down the renal tubules to be excreted in the urine. A number of theories have been proposed, and studies performed, to determine the mechanisms involved in crystal retention within the kidneys. It has been suggested that urinary transit through the nephron is too fast for crystals to grow large enough to be retained. Thus, free particle mechanism alone cannot lead to stone formation, and there must be a mechanism for crystal fixation within the kidneys. Animal model studies suggest that crystal retention is possible through both the free- and fixed-particle mechanisms. Crystal–cell interaction leads to pathological changes which promote crystal attachment to either epithelial cells or their basement membrane. Alternatively, crystals aggregate and produce large enough particles to block the tubules particularly at sites, where urinary flow is affected because of changes in the luminal diameter of the tubule. Crystal deposits plugging the openings of the ducts of Bellini may be the result of such a phenomenon. Intratubular crystals translocating to renal interstitium may produce osteogenic changes in the epithelial or endothelial cells resulting in the formation of the Randall’s plaques. Thus, fixation appears to be either through the formation of Randall’s plugs, crystal plugs clogging the openings of the ducts of Bellini or sub-epithelial crystal deposits, and the Randall’s plaques.

This is a preview of subscription content, access via your institution.


Conclusions

In the present work, we introduced a new cross-species gene expression module comparison method to make the most of animal expression data and analyze the effectiveness of animal models in drug research. Through exploring the relations between drug molecules and mouse disease models, our method was able to assess whether the corresponding model recapitulates the essential features of the human disease. If so, this model may be suitable for drug molecules screening or even to test novel therapies systematically. Moreover, through data integration, our method could mine some meaningful information for drug research, such as potential drug candidates, possible drug repositioning, side effects and information about pharmacology.


Acknowledgements

The authors are grateful to P. Bieniasz and R. Desrosiers for critically reviewing this manuscript. The authors also thank P. Marx for sharing the data included in table 1, and A. Halper-Stromberg and R. Liberatore for help with figure 1 and lessons in mouse anatomy. The authors acknowledge and regret that a number of important studies are not cited owing to space constraints. T.H. is supported by US Department of Health and Human Services Public Health Service (PHS) grants AI078788 and AI093255. D.T.E. is supported by PHS grants AI087498, AI095098, AI098485 and RR000168/OD011103, and is an Elizabeth Glaser Scientist of the Elizabeth Glaser Pediatric AIDS Foundation.


Unanswered clinical questions in heart failure

It is largely unknown whether the individual changes in gene expression and physiology that are observed in heart failure patients are adaptive or maladaptive, and how this changes with the evolution of the disease. Further questions include whether there are novel biomarkers (Lainscak and Anker, 2009), and which imaging modalities are optimal that will facilitate clinical decision-making in heart failure patients.

Two obvious deficiencies are hindering the development of new heart failure therapies: (1) high-resolution, longitudinal phenotyping of heart failure patients (i.e. at several stages in the evolution of the condition) has not yet been carried out and (2) the development of new animal models that more closely mimic the medical treatment of heart failure, and the common causes of heart failure (such as obesity and hypertension) that interact with treatment, would facilitate convergence between clinical and animal modelling fields. Given the huge recent advances in our understanding of murine biology, mouse models would arguably be of the most use to answer these questions.


A Systematic Review of Animal Models of Disuse-Induced Bone Loss

Several different animal models are used to study disuse-induced bone loss. This systematic review aims to give a comprehensive overview of the animal models of disuse-induced bone loss and provide a detailed narrative synthesis of each unique animal model.

Methods

PubMed and Embase were systematically searched for animal models of disuse from inception to November 30, 2019. In addition, Google Scholar and personal file archives were searched for relevant publications not indexed in PubMed or Embase. Two reviewers independently reviewed titles and abstracts for full-text inclusion. Data were extracted using a predefined extraction scheme to ensure standardization.

Results

1964 titles and abstracts were screened of which 653 full-text articles were included. The most common animal species used to model disuse were rats (59%) and mice (30%). Males (53%) where used in the majority of the studies and genetically modified animals accounted for 7%. Twelve different methods to induce disuse were identified. The most frequently used methods were hindlimb unloading (44%), neurectomy (15%), bandages and orthoses (15%), and botulinum toxin (9%). The median time of disuse was 21 days (quartiles: 14 days, 36 days) and the median number of animals per group subjected to disuse was 10 (quartiles: 7, 14). Random group allocation was reported in 43% of the studies. Fewer than 5% of the studies justified the number of animals per group by a sample size calculation to ensure adequate statistical power.

Conclusion

Multiple animal models of disuse-induced bone loss exist, and several species of animals have successfully been studied. The complexity of disuse-induced bone loss warrants rigid research study designs. This systematic review emphasized the need for standardization of animal disuse research and reporting.


    about animal ethics, Kantianism and utilitarianism about considerations on price elasticity and supply chain buffers about our duties to inanimate objects, in dialogue form , II and III about a Kantian consumer ethics about moral systems viewed as statistical models about charity, moral systems and moral intuition
    about objectivity and subjectivity in art about meaning in life, in art and in the works of Gerald Murnane about Helena Munktell, Jehan Alain, Harold Shapero and Arnold Rosner


Watch the video: Is it possible to establish Islam without kingdom? (October 2022).