Predicting drug polypharmacology from cell morphology readouts using variational autoencoder latent space arithmetic

A variational autoencoder (VAE) is a machine learning algorithm, useful for generating a compressed and interpretable latent space. These representations have been generated from various biomedical data types and can be used to produce realistic-looking simulated data. However, standard vanilla VAEs suffer from entangled and uninformative latent spaces, which can be mitigated using other types of VAEs such as β-VAE and MMD-VAE. In this project, we evaluated the ability of VAEs to learn cell morphology characteristics derived from cell images. We trained and evaluated these three VAE variants—Vanilla VAE, β-VAE, and MMD-VAE—on cell morphology readouts and explored the generative capacity of each model to predict compound polypharmacology (the interactions of a drug with more than one target) using an approach called latent space arithmetic (LSA). To test the generalizability of the strategy, we also trained these VAEs using gene expression data of the same compound perturbations and found that gene expression provides complementary information. We found that the β-VAE and MMD-VAE disentangle morphology signals and reveal a more interpretable latent space. We reliably simulated morphology and gene expression readouts from certain compounds thereby predicting cell states perturbed with compounds of known polypharmacology. Inferring cell state for specific drug mechanisms could aid researchers in developing and identifying targeted therapeutics and categorizing off-target effects in the future.     

Structure-guided affinity maturation of a single-chain variable fragment antibody against the Fu-bc epitope of the dengue virus envelope protein

Dengue is one of the most dominant arthropod-borne viral diseases, infecting at least 390 million people every year throughout the world. Despite this, there is no effective treatment against dengue, and the only available vaccine has already been withdrawn owing to the significant adverse effects. Therefore, passive immunotherapy using monoclonal antibodies is now being sought as a therapeutic option. To date, many dengue monoclonal antibodies have been identified, most of which are serotype-specific, and only a few of which are cross-reactive. Furthermore, antibodies that cross-react within serotypes are weakly neutralizing and frequently induce antibody-dependent enhancement, which promotes viral entry and replication. Therefore, broadly neutralizing antibodies with no risk of antibody-dependent enhancement are required for the treatment of dengue. Here, we developed a single-chain variable fragment (scFv) antibody from an anti-fusion loop E53 antibody (PDB: 2IGF). We introduced previously predicted favorable complementarity-determining region (CDR) mutations into the gene encoding the scFv antibody for affinity maturation, and the resultant variants were tested in vitro against the highly conserved fusion and bc epitope of the dengue virus envelope protein. We show some of these scFv variants with two to three substitution mutations in three different CDRs possess affinity constants (K_D) ranging from 20 to 200 nM. The scFv-mutant15, containing D31L, Y105W, and S227W substitutions, showed the lowest affinity constant, (K_D = 24 ± 7 nM), approximately 100-fold lower than its parental construct. We propose that the scFv-derivative antibody may be a good candidate for the development of an effective and safe immunotherapy.     

Evaluation of Corrosion Resistance and Characterization of Ni-Cr Coated Structure Using Plasma Spray Coating for Marine Hulls

Plasmonics - Corrosion is a natural process which gradually destructs the materials with their environment by chemical or electrochemical means. Mechanized boat hull structures used in fishing were...     

Occurrence and distribution of rhizobiophages in Indian soils.

Soil samples from 30 fields in neighbouring districts of Varanasi were screened for rhizobiophages on 60 Rhizobium strains. Plaques were observed on five strains: P1, P5, SU391 (R. leguminosarum), CB756 and 32H1 (Rhizobium sp.). Rhizobiophages infective on one or more of the five strains were present in all fields. There seems to be no correlation between presence of phage and the standing crop or the pH (7.1-8.2) of the soil. Eight distinct rhizobiophages have been isolated and characterized for host specificity, plaque morphology and maximum titer in broth.     

Relationship between NF-kappaB and trypsinogen activation in rat pancreas after supramaximal caerulein stimulation

Intra-acinar cell nuclear factor-kappaB (NF-kappaB) and trypsinogen activation are early events in secretagogue-induced acute pancreatitis. We have studied the relationship between NF-kappaB and trypsinogen activation in rat pancreas. CCK analogue caerulein induces early (within 15 min) parallel activation of both NF-kappaB and trypsinogen in pancreas in vivo as well as in pancreatic acini in vitro. However, NF-kappaB activation can be induced without trypsinogen activation by lipopolysaccharide in pancreas in vivo and by phorbol ester in pancreatic acini in vitro. Stimulation of acini with caerulein after 6 h of culture results in NF-kappaB but not trypsinogen activation. Protease inhibitors (AEBSF, TLCK, and E64d) inhibit both intracellular trypsin activity and NF-kappaB activation in caerulein stimulated acini. A chymotrypsin inhibitor (TPCK) inhibits NF-kappaB activation but not trypsin activity. The proteasome inhibitor MG-132 prevents caerulein-induced NF-kappaB activation but does not prevent trypsinogen activation. These findings indicate that although caerulein-induced NF-kappaB and trypsinogen activation are temporally closely related, they are independent events in pancreatic acinar cells. NF-kappaB activation per se is not required for the development of early acinar cell injury by supramaximal secretagogue stimulation.     

Functional domains of the ClpA and ClpX molecular chaperones identified by limited proteolysis and deletion analysis

Escherichia coli ClpA and ClpX are ATP-dependent protein unfoldases that each interact with the protease, ClpP, to promote specific protein degradation. We have used limited proteolysis and deletion analysis to probe the conformations of ClpA and ClpX and their interactions with ClpP and substrates. ATP gamma S binding stabilized ClpA and ClpX such that that cleavage by lysylendopeptidase C occurred at only two sites. Both proteins were cleaved within in a loop preceding an alpha-helix-rich C-terminal domain. Although the loop varies in size and composition in Clp ATPases, cleavage occurred within and around a conserved triad, IG(F/L). Binding of ClpP blocked this cleavage, and prior cleavage at this site rendered both ClpA and ClpX defective in binding and activating ClpP, suggesting that this site is involved in interactions with ClpP. ClpA was also cut at a site near the junction of the two ATPase domains, whereas the second cleavage site in ClpX lay between its N-terminal and ATPase domains. ClpP did not block cleavage at these other sites. The N-terminal domain of ClpX dissociated upon cleavage, and the remaining ClpXDeltaN remained as a hexamer, associated with ClpP, and expressed ATPase, chaperone, and proteolytic activity. A truncated mutant of ClpA lacking its N-terminal 153 amino acids also formed a hexamer, associated with ClpP, and expressed these activities. We propose that the N-terminal domains of ClpX and ClpA lie on the outside ring surface of the holoenzyme complexes where they contribute to substrate binding or perform a gating function affecting substrate access to other binding sites and that a loop on the opposite face of the ATPase rings stabilizes interactions with ClpP and is involved in promoting ClpP proteolytic activity.     

Activity of glutathione related enzymes and ovarian steroid hormones in different sizes of follicles from goat and sheep ovary of different reproductive stages

The investigations on enzymes related to glutathione like glutathione-S-transferase (GST) and glutathione peroxidase (GSH-Px) have been carried out mostly in human and rat ovaries, however the studies on these enzymes in ruminants are relatively absent. In the present study the changes in the activity of these enzymes, in different sizes of follicles from goat and sheep ovaries of different reproductive stages, were investigated. The results demonstrated that the activity of the enzyme GST increased with the increase in size of the follicles from small to large follicles of follicular phase ovary and from small to medium follicles of luteal phase ovary in both the species, thereafter it decreased in large follicles of luteal phase ovary. There was increasing pattern in the activity of GSH-Px in the follicular phase follicles and a decreasing pattern in the luteal phase follicles from both the species. Thus the changes in the activity of glutathione related enzymes namely GST and GSH-Px in different size follicles from both the species during different reproductive phases are evident from the results. It is reasonable, therefore, to assume that these enzymes may have functional role in the steroid hormone metabolism in ruminant ovary as reported in human ovary.