Tissue-aware interpretation of genetic variants advances the etiology of rare diseases.

Pathogenic variants underlying Mendelian diseases often disrupt the normal physiology of a few tissues and organs. However, variant effect prediction tools that aim to identify pathogenic variants are typically oblivious to tissue contexts. Here we report a machine-learning framework, denoted "Tissue Risk Assessment of Causality by Expression for variants" (TRACEvar, https://netbio.bgu.ac.il/TRACEvar/ ), that offers two advancements. First, TRACEvar predicts pathogenic variants that disrupt the normal physiology of specific tissues. This was achieved by creating 14 tissue-specific models that were trained on over 14,000 variants and combined 84 attributes of genetic variants with 495 attributes derived from tissue omics. TRACEvar outperformed 10 well-established and tissue-oblivious variant effect prediction tools. Second, the resulting models are interpretable, thereby illuminating variants' mode of action. Application of TRACEvar to variants of 52 rare-disease patients highlighted pathogenicity mechanisms and relevant disease processes. Lastly, the interpretation of all tissue models revealed that top-ranking determinants of pathogenicity included attributes of disease-affected tissues, particularly cellular process activities. Collectively, these results show that tissue contexts and interpretable machine-learning models can greatly enhance the etiology of rare diseases.

A gastrointestinal locally activating Janus kinase inhibitor to treat ulcerative colitis.

In this study, we integrated machine learning (ML), structure-tissue selectivity-activity-relationship (STAR), and wet lab synthesis/testing to design a gastrointestinal (GI) locally activating JAK inhibitor for ulcerative colitis treatment. The JAK inhibitor achieves site-specific efficacy through high local GI tissue selectivity while minimizing the requirement for JAK isoform specificity to reduce systemic toxicity. We used the ML model (CoGT) to classify whether the designed compounds were inhibitors or noninhibitors. Then we used the regression ML model (MTATFP) to predict their IC50 against related JAK isoforms of predicted JAK inhibitors. The ML model predicted MMT3-72, which was retained in the GI tract, to be a weak JAK1 inhibitor, while MMT3-72-M2, which accumulated in only GI tissues, was predicted to be an inhibitor of JAK1/2 and TYK2. ML docking methods were applied to simulate their docking poses in JAK isoforms. Application of these ML models enabled us to limit our synthetic efforts to MMT3-72 and MMT3-72-M2 for subsequent wet lab testing. The kinase assay confirmed MMT3-72 weakly inhibited JAK1, and MMT3-72-M2 inhibited JAK1/2 and TYK2. We found that MMT3-72 accumulated in the GI lumen, but not in GI tissue or plasma, but released MMT3-72-M2 accumulated in colon tissue with minimal exposure in the plasma. MMT3-72 achieved superior efficacy and reduced p-STAT3 in DSS-induced colitis. Overall, the integration of ML, the structure-tissue selectivity-activity-relationship system, and wet lab synthesis/testing could minimize the effort in the optimization of a JAK inhibitor to treat colitis. This site-specific inhibitor reduces systemic toxicity by minimizing the need for JAK isoform specificity.

Predicting molecular mechanisms of hereditary diseases by using their tissue-selective manifestation.

How do aberrations in widely expressed genes lead to tissue-selective hereditary diseases? Previous attempts to answer this question were limited to testing a few candidate mechanisms. To answer this question at a larger scale, we developed "Tissue Risk Assessment of Causality by Expression" (TRACE), a machine learning approach to predict genes that underlie tissue-selective diseases and selectivity-related features. TRACE utilized 4,744 biologically interpretable tissue-specific gene features that were inferred from heterogeneous omics datasets. Application of TRACE to 1,031 disease genes uncovered known and novel selectivity-related features, the most common of which was previously overlooked. Next, we created a catalog of tissue-associated risks for 18,927 protein-coding genes (https://netbio.bgu.ac.il/trace/). As proof-of-concept, we prioritized candidate disease genes identified in 48 rare-disease patients. TRACE ranked the verified disease gene among the patient's candidate genes significantly better than gene prioritization methods that rank by gene constraint or tissue expression. Thus, tissue selectivity combined with machine learning enhances genetic and clinical understanding of hereditary diseases.

Engineered Design of a Mesoporous Silica Nanoparticle-Based Nanocarrier for Efficient mRNA Delivery in Vivo.

We have developed tailor-designed mesoporous silica nanoparticles (MSNPs) specifically for delivering mRNA. Our unique assembly protocol involves premixing mRNA with a cationic polymer and then electrostatically binding it to the MSNP surface. Since the key physicochemical parameters of MSNPs could influence the biological outcome, we also investigated the roles of size, porosity, surface topology, and aspect ratio on the mRNA delivery. These efforts allow us to identify the best-performing carrier, which was able to achieve efficient cellular uptake and intracellular escape while delivering a luciferase mRNA in mice. The optimized carrier remained stable and active for at least 7 days after being stored at 4 °C and was able to enable tissue-specific mRNA expression, particularly in the pancreas and mesentery after intraperitoneal injection. The optimized carrier was further manufactured in a larger batch size and found to be equally efficient in delivering mRNA in mice and rats, without any obvious toxicity.

A standardized extract of Danggui Buxue Tang decoction selectively exerts estrogenic activities distinctly from tamoxifen.

More and more menopausal women use Danggui Buxue Tang (DBT) for relieving their symptoms. Concerns for its safety have been raised as it contains phytoestrogen and acts via estrogen receptors (ERs). Our study aimed to determine whether DBT could selectively exert estrogenic activities and interact with tamoxifen in bone, brain, uterus, and breast by using ovariectomized (OVX) rats and ER-positive cells. In OVX rats, DBT induced a 31.4% increase in bone mineral density and restored the mRNA expression of dopamine biomarker in striatum, 3.32-fold for tyrosine hydrolase (p < .001) and 0.21-fold for dopamine transporter (p < .001), which was similar to tamoxifen; tamoxifen, but not DBT, increased uterus weight and Complement component 3 expression by more than twofold (p < .001); unlike tamoxifen, DBT induced mild proliferation in mammary gland. Two-way ANOVA indicated the interactions between them in OVX rats (p < .05) but DBT did not alter the responses to tamoxifen. DBT stimulated proliferation or differentiation and estrogen response element in MCF-7, MG-63, Ishikawa, and SHSY5Y cells and altered the effects of tamoxifen. In summary, DBT exerted estrogenic effects in tissue-selective manner, which was different from tamoxifen. DBT interacted with tamoxifen but did not significantly alter its effects in OVX rats.

Selective debridement of burn wounds using hydrosurgery system.

In recent years, hydrosurgery is a technology that has been applied more and more in debridement procedures. However, the selectivity of hydrosurgery to cutaneous necrotic tissues has not been proved. This study was designed to investigate the possible tissue selectivity of hydrosurgery in the debridement in burn wounds. Deep partial-thickness burns were produced on the back of porcine, and 48 hours later, both burn wounds and normal skin were debrided using the hydrosurgery system. Then tissue samples were taken, and histological staining was performed and observed under microscope. Burn wound resection rates and the normal skin damaged rates were measured. Our result indicated that the burn wounds were significantly more sensitive than the normal skin when the water pressure produced by the hydrosurgery system was set between 3000 and 5000 psi (pounds per square inch), that is, the necrotic tissue portions were debrided more easily than the normal skin tissue. Based on these data, we suggest that 3000 to 5000 psi of water pressure in the hydrosurgery system has a skin tissue selectivity in burn wounds.

Cell penetration: scope and limitations by the application of cell-penetrating peptides.

The penetration of polar or badly soluble compounds through a cell membrane into live cells requires mechanical support or chemical helpers. Cell-penetrating peptides (CPPs) are very promising chemical helpers. Because of their low cytotoxicity and final degradation to amino acids, they are particularly favored in in vivo studies and for clinical applications. Clearly, the future of CPP research is bright; however, the required optimization studies for each drug require considerable individualized attention. Thus, CPPs are not the philosopher's stone. As of today, a large number of such transporter peptides with very different sequences have been identified. These have different uptake mechanisms and can transport different cargos. Intracellular concentrations of cargos can reach a low micromole range and are able to influence intracellular reactions. Internalized ribonucleic acids such as small interfering RNA (siRNA) and mimics of RNA such as peptide nucleic acids, morpholino nucleic acids, and triesters of oligonucleotides can influence transcription and translation. Despite the highly efficient internalization of antibodies, enzymes, and other protein factors, as well as siRNA and RNA mimics, the uptake and stabile insertion of DNA into the genome of the host cells remain substantially challenging. This review describes a wide array of differing CPPs, cargos, cell lines, and tissues. The application of CPPs is compared with electroporation, magnetofection, lipofection, viral vectors, dendrimers, and nanoparticles, including commercially available products. The limitations of CPPs include low cell and tissue selectivity of the first generation and the necessity for formation of fusion proteins, conjugates, or noncovalent complexes to different cargos and of cargo release from intracellular vesicles. Furthermore, the noncovalent complexes require a strong molar excess of CPPs, and extensive experimentation is required to determine the most optimal CPP for any given cargo and cell type. Yet to predict which CPP is optimal for any given target remains a complex question. More recently, there have been promising developments: the enhancement of cell specificity using activatable CPPs, specific transport into cell organelles by insertion of corresponding localization sequences, and the transport of drugs through blood-brain barriers, through the conjunctiva of eyes, skin, and into nerve cells. Proteins, siRNA, and mimics of oligonucleotides can be efficiently transported into cells and have been tested for treatment of certain diseases. The recent state of the art in CPP research is discussed together with the overall scope, limitations, and some recommendations for future research directions.

Chemical Design Strategy of Ionizable Lipids for In Vivo mRNA Delivery.

Lipid nanoparticles (LNPs) are the most clinically successful drug delivery systems that have accelerated the development of mRNA drugs and vaccines. Among various structural components of LNPs, more recent attention has been paid in ionizable lipids (ILs) that was supposed as the key component in determining the effectiveness of LNPs for in vivo mRNA delivery. ILs are typically comprised of three moieties including ionizable heads, linkers, and hydrophobic tails, which suggested that the combination of different functional groups in three moieties could produce ILs with diverse chemical structures and biological identities. In this concept article, we provide a summary of chemical design strategy for high-performing IL candidates and discuss their structure-activity relationships for shifting tissue-selective mRNA delivery. We also propose an outlook for the development of next-generation ILs, enabling the broader translation of mRNA formulated with LNPs.

Affected cell types for hundreds of Mendelian diseases revealed by analysis of human and mouse single-cell data.

Mendelian diseases tend to manifest clinically in certain tissues, yet their affected cell types typically remain elusive. Single-cell expression studies showed that overexpression of disease-associated genes may point to the affected cell types. Here, we developed a method that infers disease-affected cell types from the preferential expression of disease-associated genes in cell types (PrEDiCT). We applied PrEDiCT to single-cell expression data of six human tissues, to infer the cell types affected in Mendelian diseases. Overall, we inferred the likely affected cell types for 328 diseases. We corroborated our findings by literature text-mining, expert validation, and recapitulation in mouse corresponding tissues. Based on these findings, we explored characteristics of disease-affected cell types, showed that diseases manifesting in multiple tissues tend to affect similar cell types, and highlighted cases where gene functions could be used to refine inference. Together, these findings expand the molecular understanding of disease mechanisms and cellular vulnerability.

Drospirenone and estetrol: evaluation of a newly approved novel oral contraceptive.

INTRODUCTION: Estetrol (E4) is a native estrogen produced only by the fetal liver during pregnancy. E4 is the first new estrogen to be used in hormonal contraception since the introduction of oral contraceptives in 1960. Ethinyl estradiol, the most commonly used estrogen in oral contraceptives today, increases the risks of thromboembolism and has other significant hepatic impacts, which induce important drug-drug interactions. On the other hand, Phase 2 E4 characterization studies demonstrated that E4 has negligible impacts on liver, breast, and vascular endothelium due to its distinct tissue selectivity. Combined with drospirenone (DRSP), E4 offers an improved safety profile for oral contraception. AREAS COVERED: This paper briefly highlights the unique pharmacokinetic and pharmacodynamic features of E4. The efficacy, safety, and tolerability results from the Phase 2 and 3 studies of the E4/DRSP pill are discussed to provide the reader with a thorough understanding of E4 and information to use when counseling potential users. EXPERT OPINION: The estetrol/drospirenone oral contraceptive is effective and well tolerated and provides good cycle control. In the future, estetrol may be the estrogen of choice if subsequent evidence verifies that it reduces the risks associated with current estrogens, such as venous thromboembolism and drug-drug interactions.

Analysis of fatty acid composition and sensitivity to dietary n-3 PUFA intervention of mouse n-3 PUFA-enriched tissues/organs.

PURPOSE: n-3 polyunsaturated fatty acids (PUFAs), docosahexaenoic acid (DHA; C22:6 n3) and eicosapentaenoic acid (EPA; C20:5 n3), are of concern for their health-promoting effects such as anti-inflammatory, but the tissue selectivity for n-3 PUFA (i.e., which tissues and organs are rich in n-3 PUFA) is still not well known. In addition, it is unclear which tissues and organs are more sensitive to n-3 PUFA intervention. These unresolved issues have greatly hindered the exploring of the health benefits of n-3 PUFA. METHODS: Twenty-four 7-week-old male C57BL/6 J mice were assigned to the control, fish oil, DHA, and EPA groups. The last three groups were given a 4-week oral intervention of fatty acids in ethyl ester (400 mg/kg bw). The fatty acid profiles in 27 compartments were determined by gas chromatography. RESULTS: The proportion of long-chain n-3 PUFA (the total relative percentage of EPA, DPA n3, and DHA) was analyzed. Eight tissues and organs, including the brain (cerebral cortex, hippocampus, hypothalamus) and peripheral organs (tongue, quadriceps, gastrocnemius, kidney, and heart) were determined as being n-3 PUFA-enriched tissues and organs, owing to their high n-3 PUFA levels. The highest n-3 PUFA content was observed in the tongue for the first time. Notably, the content of linoleic acid (LA; C18:2 n6c) in peripheral organs was observed to be relatively high compared with that in the brain. Interestingly, the proportions of EPA in the kidney, heart, quadriceps, gastrocnemius, and tongue increased more markedly after the EPA intervention than after the DHA or fish oil intervention. As expected, the levels of proinflammatory arachidonic acid (AA; C20:4 n6) in the kidney, quadriceps, and tongue were markedly decreased after the three dietary interventions. CONCLUSION: Peripheral tissues and organs, including the tongue, quadriceps, gastrocnemius, kidney, and heart, besides the brain, showed obvious tissue selectivity for n-3 PUFA. In the whole body of mice, the tongue exhibits the strongest preference for n-3 PUFA, with the highest proportion of n-3 PUFA. Moreover, these peripheral tissues and organs, especially the kidney, are more sensitive to dietary EPA administration in comparison with the brain.

Estradiol and Estrogen-like Alternative Therapies in Use: The Importance of the Selective and Non-Classical Actions.

Estrogen is one of the most important female sex hormones, and is indispensable for reproduction. However, its role is much wider. Among others, due to its neuroprotective effects, estrogen protects the brain against dementia and complications of traumatic injury. Previously, it was used mainly as a therapeutic option for influencing the menstrual cycle and treating menopausal symptoms. Unfortunately, hormone replacement therapy might be associated with detrimental side effects, such as increased risk of stroke and breast cancer, raising concerns about its safety. Thus, tissue-selective and non-classical estrogen analogues have become the focus of interest. Here, we review the current knowledge about estrogen effects in a broader sense, and the possibility of using selective estrogen-receptor modulators (SERMs), selective estrogen-receptor downregulators (SERDs), phytoestrogens, and activators of non-genomic estrogen-like signaling (ANGELS) molecules as treatment.

Water Extract of Rhizoma Drynaria Selectively Exerts Estrogenic Activities in Ovariectomized Rats and Estrogen Receptor-Positive Cells.

Our previous study demonstrated that the bone protective actions of herbal medicine Rhizoma Drynariae (Gusuibu, RD) were mainly mediated by flavonoid phytoestrogens via estrogen receptors, raising concerns about the safety of using RD as it may induce estrogen-like risk-benefit profile and interact with other ER ligands, such as selective estrogen receptor modulators (SERMs), when coadministered. The present study evaluated the estrogenic activities of RD and its potential interaction with tamoxifen, a SERM, in estrogen-sensitive tissues by using mature ovariectomized (OVX) rats and ER-positive cells. Similar to but weaker than tamoxifen, RD at its clinical dose dramatically ameliorated OVX-induced changes in bone and dopamine metabolism-related markers in OVX rats. However, tamoxifen, but not RD, induced uterotrophic effects. No significant alteration in mammary gland was observed in OVX rats treated with RD, which was different from the inhibitory actions of tamoxifen. The two-way ANOVA results indicated the interactions between RD and tamoxifen in the bone, brain, and uterus of OVX rats while RD did not alter their responses to tamoxifen. Our results demonstrate that RD selectively exerts estrogenic actions in a different manner from tamoxifen. Moreover, RD interacts with tamoxifen without altering its effects in OVX rats.

Prenylflavonoid Icariin Induces Estrogen Response Element-Independent Estrogenic Responses in a Tissue-Selective Manner.

Icariin, a flavonoid phytoestrogen derived from Herba epimedii, has been reported to exert estrogenic effects in bone and activate phosphorylation of estrogen receptor (ER) α in osteoblastic cells. However, it is unclear whether icariin selectively exerts estrogenic activities in bone without inducing undesirable effects in other estrogen-sensitive tissues. The present study aimed to investigate the tissue-selective estrogenic activities of icariin in estrogen-sensitive tissues in vivo and in vitro. Long-term treatment with icariin effectively prevented bone of ovariectomized (OVX) rats from estrogen deficiency-induced osteoporotic changes in bone structure, bone mineral density, and trabecular properties. Moreover, icariin regulated the transcriptional events of estrogen-responsive genes related to bone remodeling and prevented dopaminergic neurons against OVX-induced changes by rescuing expression of estrogen-regulated tyrosine hydroxylase and dopamine transporter in the striatum. Unlike estrogen, icariin did not induce estrogenic effects in the uterus and breast in mature OVX rats or immature CD-1 mice. In vitro studies demonstrated that icariin exerted estrogen-like activities and regulated the expression of estrogen-responsive genes but did not induce estrogen response element-dependent luciferase activities in ER-positive cells. Our results support the hypothesis that icariin, through its distinct mechanism of actions in activating ER, selectively exerts estrogenic activities in different tissues and cell types.

Zinc-Chelating Small Molecules Preferentially Accumulate and Function within Pancreatic ß Cells.

Diabetes is a hyperglycemic condition characterized by pancreatic ß-cell dysfunction and depletion. Whereas methods for monitoring ß-cell function in vivo exist, methods to deliver therapeutics to ß cells are lacking. We leveraged the rare ability of ß cells to concentrate zinc to preferentially trap zinc-binding molecules within ß cells, resulting in ß-cell-targeted compound delivery. We determined that zinc-rich ß cells and islets preferentially accumulated TSQ (6-methoxy-8-p-toluenesulfonamido-quinoline) in a zinc-dependent manner compared with exocrine pancreas. Next, we asked whether appending a zinc-chelating moiety onto a ß-cell replication-inducing compound was sufficient to confer preferential ß-cell accumulation and activity. Indeed, the hybrid compound preferentially accumulated within rodent and human islets in a zinc-dependent manner and increased the selectivity of replication-promoting activity toward ß cells. These data resolve the fundamental question of whether intracellular accumulation of zinc-chelating compounds is influenced by zinc content. Furthermore, application of this principle yielded a proof-of-concept method for ß-cell-targeted drug delivery and bioactivity.

Target and Tissue Selectivity Prediction by Integrated Mechanistic Pharmacokinetic-Target Binding and Quantitative Structure Activity Modeling.

Selectivity is an important attribute of effective and safe drugs, and prediction of in vivo target and tissue selectivity would likely improve drug development success rates. However, a lack of understanding of the underlying (pharmacological) mechanisms and availability of directly applicable predictive methods complicates the prediction of selectivity. We explore the value of combining physiologically based pharmacokinetic (PBPK) modeling with quantitative structure-activity relationship (QSAR) modeling to predict the influence of the target dissociation constant (K D) and the target dissociation rate constant on target and tissue selectivity. The K D values of CB1 ligands in the ChEMBL database are predicted by QSAR random forest (RF) modeling for the CB1 receptor and known off-targets (TRPV1, mGlu5, 5-HT1a). Of these CB1 ligands, rimonabant, CP-55940, and Δ8-tetrahydrocanabinol, one of the active ingredients of cannabis, were selected for simulations of target occupancy for CB1, TRPV1, mGlu5, and 5-HT1a in three brain regions, to illustrate the principles of the combined PBPK-QSAR modeling. Our combined PBPK and target binding modeling demonstrated that the optimal values of the K D and k off for target and tissue selectivity were dependent on target concentration and tissue distribution kinetics. Interestingly, if the target concentration is high and the perfusion of the target site is low, the optimal K D value is often not the lowest K D value, suggesting that optimization towards high drug-target affinity can decrease the benefit-risk ratio. The presented integrative structure-pharmacokinetic-pharmacodynamic modeling provides an improved understanding of tissue and target selectivity.

CB1 and CB2 Receptor Pharmacology.

The CB1 and CB2 cannabinoid receptors (CB1R, CB2R) are members of the G protein-coupled receptor (GPCR) family that were identified over 20 years ago. CB1Rs and CB2Rs mediate the effects of Δ9-tetrahydrocannabinol (Δ9-THC), the principal psychoactive constituent of marijuana, and subsequently identified endogenous cannabinoids (endocannabinoids) anandamide and 2-arachidonoyl glycerol. CB1Rs and CB2Rs have both similarities and differences in their pharmacology. Both receptors recognize multiple classes of agonist and antagonist compounds and produce an array of distinct downstream effects. Natural polymorphisms and alternative splice variants may also contribute to their pharmacological diversity. As our knowledge of the distinct differences grows, we may be able to target select receptor conformations and their corresponding pharmacological responses. This chapter will discuss their pharmacological characterization, distribution, phylogeny, and signaling pathways. In addition, the effects of extended agonist exposure and how that affects signaling and expression patterns of the receptors are considered.

Proteome analysis of multidrug-resistant, breast cancer-derived microparticles.

Cancer multidrug resistance (MDR) occurs when cancer cells evade the cytotoxic actions of chemotherapeutics through the active efflux of drugs from within the cells. Our group have previously demonstrated that multidrug-resistant breast cancer cells spontaneously shed microparticles (MPs) and that these MPs can transfer resistance to drug-responsive cells and confer MDR on those cells in as little as 4 h. Furthermore, we also showed that, unlike MPs derived from leukaemia cells, breast cancer-derived MPs display a tissue selectivity in the transfer of P-glycoprotein (P-gp), transferring the resistance protein only to malignant breast cells. This study aims to define the proteome of breast cancer-derived MPs in order to understand the differences in protein profiles between those shed from drug-resistant versus drug-sensitive breast cancer cells. In doing so, we detail the protein cargo required for the intercellular transfer of MDR to drug-sensitive recipient cells and the factors governing the transfer selectivity to malignant breast cells. We describe the first proteomic analysis of MPs derived from human breast cancer cells using SDS PAGE and liquid chromatography-tandem mass spectrometry (LC/MS/MS), in which we identify 120 unique proteins found only in drug-resistant, breast cancer-derived MPs. Our results demonstrate that the MP-mediated transfer of P-gp to recipient cells occurs alongside CD44; the Ezrin, Radixin and Moesin protein family (ERM); and cytoskeleton motor proteins within the MP cargo.

Distribution of CPP-Protein Complexes in Freshly Resected Human Tissue Material.

Interest in cell-penetrating peptides (CPPs) as delivery agents has fuelled a large number of studies conducted on cultured cells and in mice. However, only a few studies have been devoted to the behaviour of CPPs in human tissues. Therefore, we performed ex vivo tissue-dipping experiments where we studied the distribution of CPP-protein complexes in samples of freshly harvested human tissue material. We used the carcinoma or hyperplasia-containing specimens of the uterus and the cervix, obtained as surgical waste from nine hysterectomies. Our aim was to evaluate the tissue of preference (epithelial versus muscular/connective tissue, carcinoma versus adjacent histologically normal tissue) for two well-studied CPPs, the transportan and the TAT-peptide. We complexed biotinylated CPPs with avidin--galactosidase (ABG), which enabled us to apply whole-mount X-gal staining as a robust detection method. Our results demonstrate that both peptides enhanced the tissue distribution of ABG. The enhancing effect of the tested CPPs was more obvious in the normal tissue and in some specimens we detected a striking selectivity of CPP-ABG complexes for the normal tissue. This unexpected finding encourages the evaluation of CPPs as local delivery agents in non-malignant situations, for example in the intrauterine gene therapy of benign gynaecological diseases.

Evaluation of two outlier-detection-based methods for detecting tissue-selective genes from microarray data.

Large-scale expression profiling using DNA microarrays enables identification of tissue-selective genes for which expression is considerably higher and/or lower in some tissues than in others. Among numerous possible methods, only two outlier-detection-based methods (an AIC-based method and Sprent's non-parametric method) can treat equally various types of selective patterns, but they produce substantially different results. We investigated the performance of these two methods for different parameter settings and for a reduced number of samples. We focused on their ability to detect selective expression patterns robustly. We applied them to public microarray data collected from 36 normal human tissue samples and analyzed the effects of both changing the parameter settings and reducing the number of samples. The AIC-based method was more robust in both cases. The findings confirm that the use of the AIC-based method in the recently proposed ROKU method for detecting tissue-selective expression patterns is correct and that Sprent's method is not suitable for ROKU.