Supplementary data are available at Bioinformatics online.

While template-free protein structure prediction protocols now produce good quality models for many targets, modelling failure remains common. For these methods to be useful it is important that users can both choose the best model from the hundreds to thousands of models that are commonly generated for a target, and determine whether this model is likely to be correct. We have developed Random Forest Quality Assessment (RFQAmodel), which assesses whether models produced by a protein structure prediction pipeline have the correct fold. RFQAmodel uses a combination of existing quality assessment scores with two predicted contact map alignment scores. These alignment scores are able to identify correct models for targets that are not otherwise captured. Our classifier was trained on a large set of protein domains that are structurally diverse and evenly balanced in terms of protein features known to have an effect on modelling success, and then tested on a second set of 244 protein domains with a similar spread of properties. When models for each target in this second set were ranked according to the RFQAmodel score, the highest-ranking model had a high-confidence RFQAmodel score for 67 modelling targets, of which 52 had the correct fold. At the other end of the scale RFQAmodel correctly predicted that for 59 targets the highest-ranked model was incorrect. In comparisons to other methods we found that RFQAmodel is better able to identify correct models for targets where only a few of the models are correct. We found that RFQAmodel achieved a similar performance on the model sets for CASP12 and CASP13 free-modelling targets. Finally, by iteratively generating models and running RFQAmodel until a model is produced that is predicted to be correct with high confidence, we demonstrate how such a protocol can be used to focus computational efforts on difficult modelling targets. RFQAmodel and the accompanying data can be downloaded from http://opig.stats.ox.ac.uk/resources.

Advancing age is the greatest risk factor for developing multiple age-related diseases. When developing therapeutics, using a Geroscience approach to target the shared underlying pathways of ageing, rather than individual diseases, may be an effective way to treat and prevent age-related morbidity while potentially reducing the burden of polypharmacy. We harness the Open Targets Platform and Open Targets Genetics Portal to perform a systematic analysis of nearly 1,400 genome-wide association studies (GWAS) mapped to 34 age-related diseases and traits to identify genetic signals that appear to be shared between two or more of these traits. We identify 995 targets with shared genetic links to these age-related diseases and traits, which are enriched in mechanisms of ageing and include known ageing and longevity-related genes. Of these 995 genes, 128 are the target of an approved or investigational drug, 526 have experimental evidence of binding pockets or are predicted to be tractable by small molecule or antibody modality approaches, and 341 have no existing tractability evidence, representing underexplored genes which may reveal novel biological insights and therapeutic opportunities. We present these candidate targets in a web application, TargetAge, to enable the exploration and prioritisation of possible novel drug targets for age-related multimorbidity.

practical application of NIS in cancer imaging and treatment through transport of 125 Iand other radioactive isotopes. Development of NIS mutants which could translocate efficiently the desired imaging ions is currently underway. Unfortunately, it is hindered by lack of understanding of the 3D organization of NIS and its relation to the anion transport. There are no known crystal structures of NIS in any of its conformational states. Instead, the structural underpinnings of NIS transport have been probed with mutagenesis studies and through parallels with other similar transporters. Homology modelling can be very effective in such situations, however the sequence identity between NIS and the relevant secondary transporters with available experimental structures is very low, which makes the choice of a template and the validation of sequence alignments and generated 3D models non-trivial.Here we report a homology model of NIS in its occluded state, based on 6 different templates (LeuT, vSGLT, Mhp1, AdiC, BetP and dDAT). Our model was validated on the basis of important mutations occurring in all 6 templates and additional proteins found in human cells (hSGLT, hSERT, hDAT). Some important binding and gating residues, which emerged from the 3D structure of the model and from preliminary molecular dynamics simulations and free energy calculations of ion-protein binding, were identified and subjected to further mutagenesis and anion uptake tests.

No abstract

While template-free protein structure prediction protocols now produce good quality models for many targets, modelling failure remains common. For these methods to be useful it is important that users can both choose the best model from the hundreds to thousands of models that are commonly generated for a target, and determine whether this model is likely to be correct. We have developed Random Forest Quality Assessment (RFQAmodel), which assesses whether models produced by a protein structure prediction pipeline have the correct fold. RFQAmodel uses a combination of existing quality assessment scores with two predicted contact map alignment scores.These alignment scores are able to identify correct models for targets that are not otherwise captured. Our classifier was trained on a large set of protein domains that are structurally diverse and evenly balanced in terms of protein features known to have an effect on modelling success, and then tested on a second set of 244 protein domains with a similar spread of properties. When models for each target in this second set were ranked according to the RFQAmodel score, the highest-ranking model had a high-confidence RFQAmodel score for 67 modelling targets, of which 52 had the correct fold. At the other end of the scale RFQAmodel correctly predicted that for 59 targets the highest-ranked model was incorrect. In comparisons to other methods we found that May 23, 2019 1/23 RFQAmodel is better able to identify correct models for targets where only a few of the models are correct. We found that RFQAmodel achieved a similar performance on the model sets for CASP12 and CASP13 free-modelling targets. Finally, by iteratively generating models and running RFQAmodel until a model is produced that is predicted to be correct with high confidence, we demonstrate how such a protocol can be used to focus computational efforts on difficult modelling targets. Introduction 1 Template-free protein structure prediction protocols routinely produce hundreds to 2 thousands of models for a given target [1]. Users need to be able to identify if a good 3 model exists in this ensemble. The final step in a typical structure prediction pipeline is 4 therefore to select a representative subset of five or fewer models as output [2]. This 5 model selection step is critical, and the community's ability to select good models is 6 assessed as part of the Critical Assessment of protein Structure Prediction (CASP) 7 experiments [3]. 8 Protocols for model quality assessment can be divided into three classes: 9 single-model methods, quasi-single model methods, and consensus methods [2]. 10 Single-model methods calculate a score for each model independently, and this score 11 does not take into account any of the other models generated for a particular target. 12 The objective function optimised during protein structure prediction can usually be 13 used as a single-model quality estimator, but better results have been reported if 14 different scores are used for modelling and ranking [2]. Examples of single...