lncLocator 2.0: a cell-line-specific subcellular localization predictor for long non-coding RNAs with interpretable deep learning

Author(s):  
Yang Lin ◽  
Xiaoyong Pan ◽  
Hong-Bin Shen

Abstract Motivation Long non-coding RNAs (lncRNAs) are generally expressed in a tissue-specific way, and subcellular localizations of lncRNAs depend on the tissues or cell lines that they are expressed. Previous computational methods for predicting subcellular localizations of lncRNAs do not take this characteristic into account, they train a unified machine learning model for pooled lncRNAs from all available cell lines. It is of importance to develop a cell-line-specific computational method to predict lncRNA locations in different cell lines. Results In this study, we present an updated cell-line-specific predictor lncLocator 2.0, which trains an end-to-end deep model per cell line, for predicting lncRNA subcellular localization from sequences.We first construct benchmark datasets of lncRNA subcellular localizations for 15 cell lines. Then we learn word embeddings using natural language models, and these learned embeddings are fed into convolutional neural network, long short-term memory and multilayer perceptron to classify subcellular localizations. lncLocator 2.0 achieves varying effectiveness for different cell lines and demonstrates the necessity of training cell-line-specific models. Furthermore, we adopt Integrated Gradients to explain the proposed model in lncLocator 2.0, and find some potential patterns that determine the subcellular localizations of lncRNAs, suggesting that the subcellular localization of lncRNAs is linked to some specific nucleotides. Availability The lncLocator 2.0 is available at www.csbio.sjtu.edu.cn/bioinf/lncLocator2 and the source code can be found at https://github.com/Yang-J-LIN/lncLocator2. Supplementary information Supplementary data are available at Bioinformatics online.

1981 ◽  
Vol 49 (1) ◽  
pp. 87-97
Author(s):  
D. Rohme

The dose response of Sendai virus-induced cell fusion was studied in 10 mammalian cell lines, comprising 5 continuous and 5 diploid cell lines originating from 5 species. The extent of fusion was calculated using a parameter directly proportional to the number of fusion events (t-parameter). At lower levels of fusion the dose response was found to be based on the same simple kinetic rules in all cell lines and was defined by the formula: t = FS. FAU/(I + FS. FAU), where FS (fusion sensitivity) is a cell-specific constant of the fusion rate and FAU (fusion activity units) is the virus dose. The FS potential of a cell line was determined as the linear regression coefficient of the fusion index (t/(I - t)) on the virus dose. At higher levels of fusion, when the fusion extent reached cell-line-specific maximal levels, the dose response was not as uniform. In general, and particularly in the cases of the diploid cell lines, these maximal levels were directly proportional to the FS potentials. Thus, it was concluded that the FS potential is the basic quantitative feature, which expresses the cellular fusion efficiency. The fact that FS varied extensively between cell lines, but at the same time apparently followed certain patterns (being higher in continuous compared to diploid cell lines and being related to the species of origin of the cells), emphasizes it biological significance as well as its possible usefulness in studies of the efficiency of various molecular interactions in the cell membrane/cytoskeleton system.


2005 ◽  
Vol 10 (8) ◽  
pp. 832-840 ◽  
Author(s):  
Heather Guthrie ◽  
Frederick S. Livingston ◽  
Ueli Gubler ◽  
Ralph Garippa

Several commercially available pharmaceutical compounds have been shown to block the I Krcurrent of the cardiac action potential. This effect can cause a prolongation of the electrocardiogram QTinterval and a delay in ventricular repolarization. The Food and Drug Administration recommends that all new potential drug candidates be assessed for I Krblock to avoid a potentially lethal cardiac arrhythmia known as torsades de pointes. Direct compound interaction with the human ether-a-go-go– related gene (hERG) product, a delayed rectifier potassium channel, has been identified as a molecular mechanism of I Kr block. One strategy to identify compounds withh ERGliability is to monitor hERGcurrent inhibition using electrophysiology techniques. The authors describe the Ion Works HT ™instrument as a tool for screening cell lines expressing hERG channels. Based on current amplitude and stability criteria, a cell line was selected and used to perform a 300-compound screen. The screen was able to identify compounds with hERG activity within projects that spanned different therapeutic areas. The cell line selection and optimization, as well as the screening abilities of the Ion Works HT ™system, provide a powerful means of assessinghERGactive compounds early in the drug discovery pipeline.


2019 ◽  
Vol 36 (5) ◽  
pp. 1607-1613 ◽  
Author(s):  
Joseph C Boyd ◽  
Alice Pinheiro ◽  
Elaine Del Nery ◽  
Fabien Reyal ◽  
Thomas Walter

Abstract Motivation High-content screening is an important tool in drug discovery and characterization. Often, high-content drug screens are performed on one single-cell line. Yet, a single-cell line cannot be thought of as a perfect disease model. Many diseases feature an important molecular heterogeneity. Consequently, a drug may be effective against one molecular subtype of a disease, but less so against another. To characterize drugs with respect to their effect not only on one cell line but on a panel of cell lines is therefore a promising strategy to streamline the drug discovery process. Results The contribution of this article is 2-fold. First, we investigate whether we can predict drug mechanism of action (MOA) at the molecular level without optimization of the MOA classes to the screen specificities. To this end, we benchmark a set of algorithms within a conventional pipeline, and evaluate their MOA prediction performance according to a statistically rigorous framework. Second, we extend this conventional pipeline to the simultaneous analysis of multiple cell lines, each manifesting potentially different morphological baselines. For this, we propose multi-task autoencoders, including a domain-adaptive model used to construct domain-invariant feature representations across cell lines. We apply these methods to a pilot screen of two triple negative breast cancer cell lines as models for two different molecular subtypes of the disease. Availability and implementation https://github.com/jcboyd/multi-cell-line or https://zenodo.org/record/2677923. Supplementary information Supplementary data are available at Bioinformatics online.


1973 ◽  
Vol 138 (2) ◽  
pp. 364-372 ◽  
Author(s):  
M. Hatanaka ◽  
R. Klein ◽  
C. W. Long ◽  
R. Gilden

Tumorigenic and nontumorigenic mutants induced by a single 5'-bromodeoxyuridine (BrdU) treatment of a nonproducer (NP) tumorigenic cell line were isolated and characterized. Among the cloned derivatives were examples of virus-free and sarcoma virus-producing cell lines. Oncogenicity did not correlate with production of virus or ease of rescue of the sarcoma genome. All lines, including nononcogenic derivatives, retained the sarcoma genome. Phenotypic reversion of some cell mutants was observed after in vivo inoculation or long term in vitro cultivation. The M-50T cell line, obtained from a tumor induced by M-50 cells, had a sarcoma genome rescuable by direct superinfection; this was only achieved with parental M-50 cells by a cell fusion rescue technique. The M-43-2T cell, obtained from a single small static tumor induced by otherwise nononcogenic M-43-2 cells, shed sarcoma virus and became tumorigenic. M-58-4-48 became tumorigenic after passage 48 of the M-58-4 line, which was originally nontumorigenic. These observations of phenotypic reversion demonstrate that the presence of the sarcoma gene in cells is an essential but not sufficient condition of tumorigenesis.


2022 ◽  
Vol 12 (1) ◽  
Author(s):  
Warren Fiskus ◽  
Steffen Boettcher ◽  
Naval Daver ◽  
Christopher P. Mill ◽  
Koji Sasaki ◽  
...  

AbstractTreatment with Menin inhibitor (MI) disrupts the interaction between Menin and MLL1 or MLL1-fusion protein (FP), inhibits HOXA9/MEIS1, induces differentiation and loss of survival of AML harboring MLL1 re-arrangement (r) and FP, or expressing mutant (mt)-NPM1. Following MI treatment, although clinical responses are common, the majority of patients with AML with MLL1-r or mt-NPM1 succumb to their disease. Pre-clinical studies presented here demonstrate that genetic knockout or degradation of Menin or treatment with the MI SNDX-50469 reduces MLL1/MLL1-FP targets, associated with MI-induced differentiation and loss of viability. MI treatment also attenuates BCL2 and CDK6 levels. Co-treatment with SNDX-50469 and BCL2 inhibitor (venetoclax), or CDK6 inhibitor (abemaciclib) induces synergistic lethality in cell lines and patient-derived AML cells harboring MLL1-r or mtNPM1. Combined therapy with SNDX-5613 and venetoclax exerts superior in vivo efficacy in a cell line or PD AML cell xenografts harboring MLL1-r or mt-NPM1. Synergy with the MI-based combinations is preserved against MLL1-r AML cells expressing FLT3 mutation, also CRISPR-edited to introduce mtTP53. These findings highlight the promise of clinically testing these MI-based combinations against AML harboring MLL1-r or mtNPM1.


2014 ◽  
Author(s):  
Jee-Hyub Kim

A cell line is a cell culture developed from a single cell and therefore consisting of cells with a uniform genetic make-up. A cell line has an important role as a research resource such as organisms, antibodies, constructs, knockdown reagents, etc. Unique identification of cell lines in the biomedical literature is important for the reproducibility of science. As data citation, resource citation is also important for resource re-use. In this paper, we mention the challenges of identifying cell lines and describe a system for cell line annotation with preliminary results.


2019 ◽  
Vol 93 (13) ◽  
Author(s):  
Diem-Lan Vu ◽  
Albert Bosch ◽  
Rosa M. Pintó ◽  
Enric Ribes ◽  
Susana Guix

ABSTRACT MLB astroviruses were identified 10 years ago in feces from children with gastroenteritis of unknown etiology and have been unexpectedly detected in severe cases of meningitis/encephalitis, febrile illness of unknown etiology, and respiratory syndromes. The aim of this study was to establish a cell culture system supporting MLB astrovirus replication. We used two clinical strains to infect several cell lines, an MLB1 strain from a gastroenteritis case, and an MLB2 strain associated with a neurologic infection. Efforts to propagate the viruses in the Caco-2 cell line were unsuccessful. In contrast, we identified two human nonintestinal cell lines, Huh-7 and A549, permissive for both genotypes. After serial passages in the Huh-7.5 cell line, the adapted strains were able to establish persistent infections in the Huh-7.5, Huh-7AI, and A549 cell lines, with high viral loads (up to 10 log10 genome copies/ml) detected by quantitative reverse transcription-PCR (RT-qPCR) in the culture supernatant. Immunofluorescence assays demonstrated infection in about 10% of cells in persistently infected cultures. Electron microscopy revealed particles of 32 to 33 nm in diameter after negative staining of cell supernatants and capsid arrays in ultrathin sections with a particularly high production in Huh-7.5 cells. Interferon (IFN) expression by infected cells and effect of exogenous IFN varied depending on the type of infection and the cell line. The availability of a cell culture system to propagate MLB astroviruses represents a key step to better understand their replicative cycle, as well as a source of viruses to conduct a wide variety of basic virologic studies. IMPORTANCE MLB astroviruses are emerging viruses infecting humans. More studies are required to determine their exact epidemiology, but several reports have already identified them as the cause of unexpected clinical diseases, including severe neurologic diseases. Our study provides the first description of a cell culture system for the propagation of MLB astroviruses, enabling the study of their replicative cycle. Moreover, we demonstrated the unknown capacity of MLB astrovirus to establish persistent infections in cell culture. Whether these persistent infections are also established in vivo remains unknown, but the clinical consequences would be of high interest if persistence was confirmed in vivo. Finally, our analysis of IFN expression provides some trails to understand the mechanism by which MLB astroviruses can cause persistent infections in the assayed cultures.


2019 ◽  
Author(s):  
Ahmed Ibrahim Samir Khalil ◽  
Anupam Chattopadhyay ◽  
Amartya Sanyal

Abstract Background The widespread concern about genetic drift and cross-contamination of cell lines calls for a pressing need for their authentication. The current genetic techniques for authentication are time-consuming and require specific documentary standard and laboratory protocols. Given the fact that whole-genome sequencing (WGS) data are readily available, read depth (RD)-based computational analyses has allowed the estimation of genetic profiles of cell lines. Results We propose WGS-derived aneuploidy profiling as a prototype of digital karyotyping for authentication of cancer cell lines. Here, we describe a Python-based software AStra for de novo estimation of the genome-wide aneuploidy profile, the copy number of every genomic loci, from raw WGS reads. We demonstrated that aneuploidy profile offers a unique signature that can distinguish the clonal variants (strains) of a cell line. We evaluated our approach using simulated data and variety of cancer cell lines. We further showed that cell lines exhibit distinct aneuploidy patterns which corroborate well with the experimental observations. Conclusions AStra is a simple, user-friendly, and free tool that provides the elementary information about the chromosomal aneuploidy for cell line authentication. AStra provides an analytical and visualization platform for rapid and easy comparison between different cell lines/strains. We recommend AStra for rapid first-pass quality assessment of scientific data that employ cancer cell lines. AStra is an open source software and is available at https://github.com/AISKhalil/AStra.


Blood ◽  
2016 ◽  
Vol 128 (22) ◽  
pp. 2950-2950
Author(s):  
Shohei Mizuno ◽  
Ichiro Hanamura ◽  
Akinobu Ota ◽  
Sivasundaram Karnan ◽  
Kaori Uchino ◽  
...  

Abstract Introduction: Primary refractory DLBCL is an extremely difficult condition to treat and represent an unmet medical need. The clarification of the molecular pathogenesis of this disease can contribute the development of new therapeutic possibilities. PVT1 is located adjacent to MYC at 8q24 and a non-protein coding gene. PVT1 produces a variety of long non-coding RNAs (lncRNAs) and has now emerged as a potential regulator in the pathogenesis of a lot of cancers. However, the precise biological and clinical significance of PVT1 remains largely unknown. In this study we established a novel human DLBCL cell line with hsr (homogeneous staining region) of 8q24 inducing high expression of PVT1 lncRNAs, named AMU-ML2. The cell line was established from a patient with primary refractory DLBCL before initiation of chemotherapy. We analyzed the genetic characteristics and investigated the drug sensitivity of the cell line in comparison with other B-cell lymphoma cell lines which had 8q24 abnormalities. Case: A 64-year-old man was diagnosed as DLBCL involved with bone marrow and pleural effusion. He was refractory to initial R-CHOP and subsequent R-hyper-CVAD/MA therapy and died of Trichosporon asahiisepsis 6 months after diagnosis. His lymphoma cells at diagnosis were positive for CD19, CD20, BCL6 and HLA-DR, and negative for CD3, CD5, CD10, cyclinD1, BCL2, MUM1, TP53 and EBER by flow cytometry and/or immunohistochemical staining, indicative of germinal center B-cell-like (GCB) DLBCL. The representative karyotype of cells was 46,XY, including add(6)(p11), add(8)(q24), hsr 8q24, add(9)(p13) and add(17)(p11.2). Establishment of the cell line: After 2 weeks of culture, the cells in pleural effusion collected before chemotherapy started to grow in suspension. The cell line was designated as AMU-ML2 after confirmation that the cells started growing again after the conventional freeze-thaw procedure. Materials and methods: B-cell lymphoma cell lines; AMU-ML2, SU-DHL-10, Raji, P3HR-1 and VAL were used in the present study. These cell lines had t(8;14)(q24;q32) or 8q24 amplification. The genomic aberration of AMU-ML2 was analyzed by cytogenetics including G-banding and FISH (fluorescence in situ hybridization) combined with array-CGH (SurePrint G3, Agilent). The TaqMan real time RT-PCR was used for measurement of expression levels of MYC and PVT1. MTT assay was used for the cell proliferation to analyze the drug sensitivities of cyclophosphamide, doxorubicin, vincristine (VCR) and prednisolone. Results: AMU-ML2 cells showed same immunophenotypic feature and karyotype as the primary sample from the patient. FISH using a MYC/IGH probe set showed no fusion signal for IGH and MYC; however, the MYC copy number was extremely increased, corresponding to hsr on chromosome 8q24. Array-CGH revealed that a 1,462 kb region, containing both the entire MYC and PVT1 genes at 8q24.21, was amplified, with greater than 20 copies present in cells (Figure 1). In addition to amplification of the MYC/PVT1 region at 8q24, 14 additional copy number alterations were detected. These included segment losses on 6p22.1-6p21.31 and 17p13 led to the LOH of human leukocyte antigen (HLA) loci and TP53, respectively (Figure 1). Using the real time RT-PCR, the expression level of the PVT1 lncRNAs were highest in AMU-ML2 among other B-cell lymphoma cell lines that we used, while the expression level of MYC in AMU-ML2 was relatively low. The proliferation rate of AMU-ML2 was significantly higher after 72-h exposure to VCR (100, 500 and 1,000 nM) compared with other cell lines (Figure 2). Discussion: AMU-ML2 was established in 2-week culture from a refractory patient before starting the chemotherapy, therefore this cell line seems to reflect the real nature of primary refractory DLBCL, not related to chemotherapy and/or cell culture. MYC amplification and the LOH of TP53 and HLA may contribute to the development of lymphoma in our case. High expression of PVT1 lncRNAs was likely to be more related with drug resistance to VCR than MYC expression, although AMU-ML2 had co-amplification of MYC and PVT1. Thus, AMU-ML2 can provide insight into the genetic and biological features and the therapeutic approaches in primary refractory DLBCL. Disclosures Ueda: Mundipharma KK: Consultancy; Kyowa Hakko Kirin: Research Funding.


Sign in / Sign up

Export Citation Format

Share Document