A comparative study of cell classifiers for image-based high-throughput screening

Background

Millions of cells are present in thousands of images created in high-throughput screening (HTS). Biologists could classify each of these cells into a phenotype by visual inspection. But in the presence of millions of cells this visual classification task becomes infeasible. Biologists train classification models on a few thousand visually classified example cells and iteratively improve the training data by visual inspection of the important misclassified phenotypes. Classification methods differ in performance and performance evaluation time. We present a comparative study of computational performance of gentle boosting, joint boosting CellProfiler Analyst (CPA), support vector machines (linear and radial basis function) and linear discriminant analysis (LDA) on two data sets of HT29 and HeLa cancer cells.

Results

For the HT29 data set we find that gentle boosting, SVM (linear) and SVM (RBF) are close in performance but SVM (linear) is faster than gentle boosting and SVM (RBF). For the HT29 data set the average performance difference between SVM (RBF) and SVM (linear) is 0.42 %. For the HeLa data set we find that SVM (RBF) outperforms other classification methods and is on average 1.41 % better in performance than SVM (linear).

Conclusions

Our study proposes SVM (linear) for iterative improvement of the training data and SVM (RBF) for the final classifier to classify all unlabeled cells in the whole data set.

Electronic supplementary material

The online version of this article (doi:10.1186/1471-2105-15-342) contains supplementary material, which is available to authorized users.     

Reconstitution of Uracil DNA Glycosylase-initiated Base Excision Repair in Herpes Simplex Virus-1

Herpes simplex virus-1 is a large double-stranded DNA virus that is self-sufficient in a number of genome transactions. Hence, the virus encodes its own DNA replication apparatus and is capable of mediating recombination reactions. We recently reported that the catalytic subunit of the HSV-1 DNA polymerase (UL30) exhibits apurinic/apyrimidinic and 5′-deoxyribose phosphate lyase activities that are integral to base excision repair. Base excision repair is required to maintain genome stability as a means to counter the accumulation of unusual bases and to protect from the loss of DNA bases. Here we have reconstituted a system with purified HSV-1 and human proteins that perform all the steps of uracil DNA glycosylase-initiated base excision repair. In this system nucleotide incorporation is dependent on the HSV-1 uracil DNA glycosylase (UL2), human AP endonuclease, and the HSV-1 DNA polymerase. Completion of base excision repair can be mediated by T4 DNA ligase as well as human DNA ligase I or ligase IIIα-XRCC1 complex. Of these, ligase IIIα-XRCC1 is the most efficient. Moreover, ligase IIIα-XRCC1 confers specificity onto the reaction in as much as it allows ligation to occur in the presence of the HSV-1 DNA polymerase processivity factor (UL42) and prevents base excision repair from occurring with heterologous DNA polymerases. Completion of base excision repair in this system is also dependent on the incorporation of the correct nucleotide. These findings demonstrate that the HSV-1 proteins in combination with cellular factors that are not encoded by the virus are capable of performing base excision repair. These results have implications on the role of base excision repair in viral genome maintenance during lytic replication and reactivation from latency.Herpes simplex virus-1 (HSV-1)² is a large double-stranded DNA virus with a genome of ∼152 kilobase pairs (for reviews, see Refs. 1 and 2). HSV-1 switches between lytic replication in epithelial cells and a state of latency in sensory neurons during which there is no detectable DNA replication (1). Viral DNA replication is mediated by seven essential virus-encoded factors (3–5). Of these, two encode subunits of the viral replicase (for review, see Refs. 6 and 7). The catalytic subunit (UL30) exhibits DNA polymerase (Pol), 3′-5′ proofreading exonuclease, and RNase H activities (8–11). UL30 exists as a heterodimer with the UL42 protein that confers a high degree of processivity on the Pol (11–17).Viral DNA replication is accompanied by vigorous recombination that leads to the formation of large networks of viral DNA replication intermediates (18). The HSV-1 single-strand DNA-binding protein (ICP8) has been shown to play a major role in mediating these recombination reactions (19–21). One role for the high frequency of recombination is to restart DNA replication at sites of fork collapse. Further mechanisms that contribute to genome maintenance are processes that survey and repair damage to the DNA to ensure the availability of a robust replication template. In this regard base excision repair (BER) is essential to remove unusual bases from the DNA and to repair apurinic/apyrimidinic (AP) sites resulting from spontaneous base loss (for review, see Ref. 22). With respect to HSV-1, a recent study showed that viral DNA from infected cultured fibroblasts contains a steady state of 2.8–5.9 AP sites per viral genome equivalent (23). Because AP sites are non-instructional, the failure to repair such sites would terminate viral replication. Indeed, UL30 cannot replicate beyond a model AP site (tetrahydrofuran residue) (23), indicating that the virus must enable a process to repair such lesions. In this regard HSV-1 possesses several enzymes that would safeguard from the accumulation of unusual bases, specifically uracil, and base loss. Hence, HSV-1 encodes a uracil DNA glycosylase (UDG) (UL2) as well as a dUTPase to reduce the pool of dUTP and prevent misincorporation by the viral Pol (24, 25). Moreover, we recently showed that the catalytic subunit of the viral Pol (UL30) exhibits AP and 5′-deoxyribose phosphate (dRP) lyase activities (26). The presence of a virus-encoded UDG and DNA lyase indicates that HSV-1 has the capacity to perform integral steps of BER, specifically for the removal of uracil. Indeed, the excision of uracil may be important for viral replication. Hence, it has been shown that uracil substitutions in the viral origins of replication alters their recognition by the viral initiator protein (27). Moreover, whereas UL2 may be dispensable for viral replication in fibroblast (24), UL2 mutants exhibit reduced neurovirulence and a decreased frequency of reactivation from latency (28). Thus, UDG action in HSV-1 may be important for viral reactivation after quiescence in neuronal cells during which the genome may accumulate uracil as a result of spontaneous deamination of cytosine. In another herpesvirus, cytomegalovirus, the viral UDG was shown to be required for the transition to late-phase DNA replication (29, 30). Consequently, it is possible that BER plays a significant role in various aspects of the herpesvirus life cycle.In mammalian single-nucleotide BER initiated by monofunctional DNA glycosylases, the resulting AP sites are incised hydrolytically at the 5′ side by AP endonuclease (APE), generating a 3′-OH. This is followed by template-directed incorporation of one nucleotide by Pol β to generate a 5′-dRP flap (22, 31, 32). The 5′-dRP residue is subsequently removed by the 5′-dRP lyase activity of Pol β to leave a nick with a 3′-OH and 5′-phosphate that is ligated by DNA ligase I or the physiologically more relevant ligase IIIα-XRCC1 complex (for review, see Refs. 33 and 34). Here we show that the HSV-1 UDG (UL2) and Pol (UL30) cooperate with human APE and human ligase IIIα-XRCC1 complex to perform BER in vitro. This finding has implications on the role of BER in viral genome maintenance during lytic replication and in the emergence of the virus from neuronal latency.     

In Silico Characterization and Homology Modeling of Thylakoidbound Ascorbate Peroxidase from a Drought Tolerant Wheat Cultivar

Ascorbate peroxidase,a haem protein (EC 1.11.1.11),efficiently scavenges hydrogen peroxide (H2O2) in cytosol and chloroplasts of plants.In this study,a fulllength coding sequence of thylakoid-bound ascorbate peroxidase cDNA (TatAPX) was cloned from a drought tolerant wheat cultivar C306.Homology modeling of the TatAPX protein was performed by using the template crystal structure of chloroplastic ascorbate peroxidase from tobacco plant (PDB: 1IYN).The model structure was further refined by molecular mechanic...     

Conformational organizations of G-quadruplexes composed of d(G4Tn)3G4

Structural polymorphism is one of the important issues with regard to G-quadruplexes because the structural diversity may significantly affect their biological functions in vivo and their physical property in nano-material. A series of oligonucleotides with four repeat guanines sequence [d(G₄T_n)₃G₄ (n = 1–6)] were designed. In this study, the effects of loop length on the formation of structures of G-quadruplex were investigated through the result of CD (circular dichroism) and 20% non-denatured polyacrylamide gel electrophoresis. Our studies demonstrate that the length of loop in 100 mM KCl solution could predict the conformation of G-quadruplex.