Using deep learning to annotate the protein universe.

Understanding the relationship between amino acid sequence and protein function is a long-standing challenge with far-reaching scientific and translational implications. State-of-the-art alignment-based techniques cannot predict function for one-third of microbial protein sequences, hampering our ability to exploit data from diverse organisms. Here, we train deep learning models to accurately predict functional annotations for unaligned amino acid sequences across rigorous benchmark assessments built from the 17,929 families of the protein families database Pfam. The models infer known patterns of evolutionary substitutions and learn representations that accurately cluster sequences from unseen families. Combining deep models with existing methods significantly improves remote homology detection, suggesting that the deep models learn complementary information. This approach extends the coverage of Pfam by >9.5%, exceeding additions made over the last decade, and predicts function for 360 human reference proteome proteins with no previous Pfam annotation. These results suggest that deep learning models will be a core component of future protein annotation tools.

Large-scale machine-learning-based phenotyping significantly improves genomic discovery for optic nerve head morphology.

Genome-wide association studies (GWASs) require accurate cohort phenotyping, but expert labeling can be costly, time intensive, and variable. Here, we develop a machine learning (ML) model to predict glaucomatous optic nerve head features from color fundus photographs. We used the model to predict vertical cup-to-disc ratio (VCDR), a diagnostic parameter and cardinal endophenotype for glaucoma, in 65,680 Europeans in the UK Biobank (UKB). A GWAS of ML-based VCDR identified 299 independent genome-wide significant (GWS; p ≤ 5 × 10-8) hits in 156 loci. The ML-based GWAS replicated 62 of 65 GWS loci from a recent VCDR GWAS in the UKB for which two ophthalmologists manually labeled images for 67,040 Europeans. The ML-based GWAS also identified 93 novel loci, significantly expanding our understanding of the genetic etiologies of glaucoma and VCDR. Pathway analyses support the biological significance of the novel hits to VCDR: select loci near genes involved in neuronal and synaptic biology or harboring variants are known to cause severe Mendelian ophthalmic disease. Finally, the ML-based GWAS results significantly improve polygenic prediction of VCDR and primary open-angle glaucoma in the independent EPIC-Norfolk cohort.

Rapid Prediction of Electron-Ionization Mass Spectrometry Using Neural Networks.

When confronted with a substance of unknown identity, researchers often perform mass spectrometry on the sample and compare the observed spectrum to a library of previously collected spectra to identify the molecule. While popular, this approach will fail to identify molecules that are not in the existing library. In response, we propose to improve the library's coverage by augmenting it with synthetic spectra that are predicted from candidate molecules using machine learning. We contribute a lightweight neural network model that quickly predicts mass spectra for small molecules, averaging 5 ms per molecule with a recall-at-10 accuracy of 91.8%. Achieving high-accuracy predictions requires a novel neural network architecture that is designed to capture typical fragmentation patterns from electron ionization. We analyze the effects of our modeling innovations on library matching performance and compare our models to prior machine-learning-based work on spectrum prediction.

Running performance and thermal sensation in the heat are improved with menthol mouth rinse but not ice slurry ingestion.

The purpose of this study was to compare the effects of a cooling strategy designed to predominately lower thermal state with a strategy designed to lower thermal sensation on endurance running performance and physiology in the heat. Eleven moderately trained male runners completed familiarization and three randomized, crossover 5-km running time trials on a non-motorized treadmill in hot conditions (33 °C). The trials included ice slurry ingestion before exercise (ICE), menthol mouth rinse during exercise (MEN), and no intervention (CON). Running performance was significantly improved with MEN (25.3 ± 3.5 min; P = 0.01), but not ICE (26.3 ± 3.2 min; P = 0.45) when compared with CON (26.0 ± 3.4 min). Rectal temperature was significantly decreased with ICE (by 0.3 ± 0.2 °C; P < 0.01), which persisted for 2 km of the run and MEN significantly decreased perceived thermal sensation (between 4 and 5 km) and ventilation (between 1 and 2 km) during the time trial. End-exercise blood prolactin concentration was elevated with MEN compared with CON (by 25.1 ± 24.4 ng/mL; P = 0.02). The data demonstrate that a change in the perception of thermal sensation during exercise from menthol mouth rinse was associated with improved endurance running performance in the heat. Ice slurry ingestion reduced core temperature but did not decrease thermal sensation during exercise or improve running performance.

The Reliability of Running Performance in a 5 km Time Trial on a Non-motorized Treadmill.

The purpose of the study was to establish the reliability of performance and physiological responses during a self-paced 5 km running time trial on a non-motorized treadmill. 17 male runners (age: 32±13 years, height: 177±7 cm, body mass: 71±9 kg, sum of 7 skinfolds: 55±21 mm) performed familiarization then 2 separate maximal 5 km running time trials on a non-motorized treadmill. Physiological responses measured included heart rate, oxygen uptake, expired air volume, blood lactate concentration, tissue saturation index and integrated electromyography. Running time (1,522±163 s vs. 1,519±162 s for trials 1 and 2, respectively) demonstrated a low CV of 1.2% and high ICC of 0.99. All physiological variables had CVs of less than 4% and ICCs of >0.92, with the exception of blood lactate concentration (7.0±2 mmol·L(-1) vs. 6.5±1.5 mmol·L(-1) for trials 1 and 2, respectively; CV: 12%, ICC: 0.83) and the electromyography measures (CV: 8-27%, ICC: 0.71-0.91). The data demonstrate that performance time in a 5 km running time trial on a non-motorized treadmill is a highly reliable test. Most physiological responses measured across the 5 km run also demonstrated good reliability.

Periodontal disease: modulation of the inflammatory cascade by dietary n-3 polyunsaturated fatty acids.

Periodontal disease, including gingivitis and periodontitis, is caused by the interaction between pathogenic bacteria and the host immune system. The ensuing oxidative stress and inflammatory cascade result in the destruction of gingival tissue, alveolar bone and periodontal ligament. This article reviews the underlying mechanisms and host-bacteria interactions responsible for periodontal disease and evidence that nutritional supplementation with fish oil may provide a protective effect. Historical investigations of diet and disease have highlighted an inverse relationship between ingestion of fish oil, which is high in n-3 polyunsaturated fatty acids, and the incidence of typical inflammatory diseases such as arthritis and coronary heart disease. Ingestion of n-3 polyunsaturated fatty acids, such as docosahexaenoic acid and eicosapentaenoic acid, results in their incorporation into membrane phospholipids, which can alter eicosanoid production after stimulation during the immune response. These eicosanoids promote a reduction in chronic inflammation, which has led to the proposal that fish oil is a possible modulator of inflammation and may reduce the severity of periodontal diseases. Tentative animal and human studies have provided an indication of this effect. Further human investigation is needed to establish the protective effects of fish oil in relation to periodontal disease.

Proteinuria in aging rats due to low-protein diet during mid-gestation.

Nephrogenesis in the rat starts mid-gestation and continues into lactation. Maternal low protein (LP) intake leads to renal injury in rats and associates with mild renal injury in humans. We hypothesized that LP during early nephrogenesis or throughout gestation would induce more renal injury in rat offspring than when LP was only present before nephrogenesis. Pregnant rats were fed LP diet (9% casein) at early gestation (LPE, day 0-7), mid (LPM, day 8-14), late (LPL, day 15-22) or throughout gestation (LPA, day 0-22) and compared to controls on 18% casein diet. Offspring were studied at 18 months. Renal injury was assessed by 24 h proteinuria, plasma urea, antioxidant enzyme activities, and apoptosis (Bax/Bcl2). Proteinuria was higher in LPM males and LPE and LPM females. In LPM males glutathione peroxidase activity was lower, while in LPE males catalase activity was higher. Antioxidants were not much affected in females. Bax expression was higher in LPM males and females, while Bcl2 expression was higher in LPA females. Thus even before nephrogenesis (day 0-7), LP impacted on renal integrity in adult life, while LP during a later phase (day 15-22) or throughout gestation had less effect. In summary, for aging rat kidney LP poses the greatest threat when restricted to early nephrogenesis.

Exposure to undernutrition in fetal life determines fat distribution, locomotor activity and food intake in ageing rats.

OBJECTIVE: To assess the long-term impact of undernutrition during specific periods of fetal life, upon central adiposity, control of feeding behaviour and locomotor activity. DESIGN: Pregnant rats were fed a control or low-protein (LP) diet, targeted to early (LPE), mid (LPM) or late (LPL) pregnancy or throughout gestation (LPA). The offspring were studied at 9 and 18 months of age. MEASUREMENTS: Adiposity was assessed by measuring weight of abdominal fat depots relative to body weight. Locomotor activity was assessed using an infrared sensor array system in both light and dark conditions. Hypothalamic expression of mRNA for galanin and the galanin 2 receptor (Gal2R) was determined using real-time PCR. RESULTS: At 9 months, male rats exposed to LP in utero had less fat in the gonadal depot, but were of similar body weight to controls. By 18 months, the males of groups LPA and LPM had more abdominal and less subcutaneous fat. Females deposited more fat centrally than males between 9 and 18 months of age, and this was more marked in groups LPA and LPL. Food intake was greater in LPM males. Among females hypophagia was noted in groups LPA and LPL. Expression of galanin and Gal2R were unaffected by maternal diet. Total locomotor activity was reduced in LPE males and all LP females in the light but not in the dark. CONCLUSION: Locomotor activity and feeding behaviour in aged rats are subject to prenatal programming influences. Fetal undernutrition does not programme obesity in rats without postnatal dietary challenge.

Cloning and partial nucleotide sequence of Schistosoma japonicum paramyosin: a potential vaccine candidate against schistosomiasis.

Paramyosin from the blood fluke, Schistosoma mansoni, has shown promise as a vaccine candidate for schistosomiasis mansoni. Here we report the cloning and partial nucleotide sequence of a cDNA encoding paramyosin from the related human parasite, Schistosoma japonicum. Affinity purified antibodies to this clone recognized a S. japonicum antigen of molecular weight 97 kDa, equivalent to the reported size of S. mansoni paramyosin. Alignment of the cDNA sequence with that of S. mansoni paramyosin revealed 90% identity. Comparison of the predicted amino acid sequences revealed 95% identity. Although these two parasites differ in many characteristics, the substantial homology demonstrated here between S. mansoni and S. japonicum paramyosin could have important implications for the development of a S. japonicum vaccine.

A review of the molecular basis of hypoxanthine-guanine phosphoribosyltransferase (HPRT) deficiency.

Hypoxanthine-guanine phosphoribosyltransferase (HPRT, EC 2.4.2.8) is a purine salvage enzyme that catalyses the conversion of hypoxanthine and guanine to their respective mononucleotides. Partial deficiency of this enzyme can result in the overproduction of uric acid leading to a severe form of gout, whilst a virtual absence of HPRT activity causes the Lesch-Nyhan syndrome which is characterised by hyperuricaemia, mental retardation, choreoathetosis and compulsive self-mutilation. The HPRT-encoding gene is located on the X chromosome in the region q26-q27 and consists of nine exons and eight introns totalling 57 kb. This gene is transcribed to produce an mRNA of 1.6 kb, which contains a protein encoding region of 654 nucleotides. With the advent of increasingly refined techniques of molecular biology, it has been possible to study the HPRT gene of individuals with a deficiency in HPRT activity to determine the genetic basis of the enzyme deficiency. Many different mutations throughout the coding region have been described, but in the absence of precise information on the three-dimensional structure of the HPRT protein, it remains difficult to determine any consistent correlation between the structure and function of the enzyme.

The molecular characterisation of HPRT CHERMSIDE and HPRT COORPAROO: two Lesch-Nyhan patients with reduced amounts of mRNA.

A complete deficiency of the purine salvage enzyme, hypoxanthine phosphoribosyltransferase (HPRT; EC 2.4.2.8), in man results in the Lesch-Nyhan (LN) syndrome. Two unrelated patients with the full LN syndrome showed no evidence of a major alteration to the gene encoding HPRT (HPRT) by restriction endonuclease analysis, but exhibited negligible levels of HPRT mRNA on Northern blots. DNA from these patients was characterised further. Amplification, by the polymerase chain reaction (PCR), of individual HPRT-exon fragments from genomic DNA followed by nucleotide (nt) sequence analysis using automated technology, revealed single-base mutations in each patient. One patient has an insertion of a T within exon-2, which places a stop codon in frame, presumably resulting in premature termination of translation of the HPRT mRNA. The other patient has a G----A base substitution at the 5' end of intron-6, at the junction of exon-6 and intron-6. Although dot blot analysis indicated negligible HPRT mRNA in lymphoblast cells from both patients, we were successful in amplifying HPRT cDNA using PCR. Direct nt sequence analysis of the amplified cDNA confirmed the insertion of a T in exon-2 in the one patient and revealed a complete deletion of exon-6 in the other patient, the latter event presumably arising due to aberrant splicing of primary message. Both mutations were also confirmed by hybridisation of amplified genomic DNA with allele-specific oligodeoxyribonucleotide probes. This study illustrates two approaches for analysing DNA mutations at the molecular level and demonstrates the power of PCR technology in the study of genetic diseases.(ABSTRACT TRUNCATED AT 250 WORDS)

Hypoxanthine-guanine phosphoribosyltransferase deficiency: analysis of HPRT mutations by direct sequencing and allele-specific amplification.

The Lesch-Nyhan syndrome is a severe X chromosome-linked human disease caused by a virtual absence of hypoxanthine-guanine phosphoribosyltransferase (HPRT) activity. A partial deficiency in the activity of this enzyme can result in gouty arthritis. To determine the genetic basis for reduction or loss of enzyme activity, we have amplified and sequenced the coding region of HPRT cDNA from four patients: one with Lesch-Nyhan syndrome (HPRTPerth) and three with partial deficiencies of HPRT activity, which have been designated HPRTUrangan, HPRTSwan and HPRTToowong. In all four patients, the only mutation identified was a single base substitution in exons 2 or 3 of the coding region, which in each case predicts a single amino acid substitution in the translated protein. Each base change was confirmed by allele-specific amplification of the patient's genomic DNA. It is interesting to note that the mutation found for HPRTPerth is identical to that reported for HPRTFlint. It appears that the two mutations are de novo events.

Identification of a single nucleotide substitution in the coding sequence of in vitro amplified cDNA from a patient with partial HPRT deficiency (HPRTBRISBANE).

The change in genomic DNA responsible for HPRT deficiency has been determined in a patient with urate overproduction and gout. In erythrocyte cell lysates, this patient had approximately 10% of normal HPRT enzyme activity and 26% of immunoidentical HPRT protein. Cultured lymphoblasts derived from this patient were used to extract mRNA. This was reverse transcribed to cDNA, which was then amplified using the polymerase chain reaction. The resulting DNA was cloned and the nucleotide sequence determined. In addition a portion of the sequence was derived from cloned double-stranded cDNA prepared by conventional first and second strand synthesis. A single nucleotide base change (a C----T transition) was detected, which predicts an amino acid substitution of isoleucine for threonine at amino acid 168 of the HPRT protein. The nucleotide substitution creates a BamHI site, confirming a restriction fragment length polymorphism previously reported in this patient.

Characterization of genomic DNA, mRNA and enzyme protein in cases of HPRT-deficiency.

1. Immunological quantitation of the HPRT proteins, together with DNA and RNA studies have defined further the heterogeneous nature of HPRT-deficiency in our patients. 2. These studies have dictated possible approaches for further characterisation of the HPRT enzyme in our patients. In the two Lesch-Nyhan patients, both the protein and the usual cDNA approach would appear difficult. 3. A BamHI polymorphism has been detected in Patient A. 4. Sequence data confirmed the creation of this BamHI site by a single C----T transition at position 602 in the coding sequence. 5. Sequencing of other patients is proceeding and use is being made of the Polymerase Chain Reaction (PCR)10 for amplification of specific segments of HPRT coding sequence.

Epstein-Barr virus nuclear antigens 1 and 2 in Burkitt lymphoma cell lines containing either 'A'- or 'B'-type virus.

Epstein-Barr virus (EBV) has previously been classified into two different types according to the organization of the EB nuclear antigen 2 (EBNA2) gene region. Type A virus hybridizes with probes from B95-8 or M-ABA viruses and the B type virus with probes from the Jijoye virus strain. The substituted region in EBV type B codes for a different, but related EBNA2 antigen, named EBNA2B as opposed to EBNA2A. In this study Burkitt lymphoma cell lines, previously typed according to the EBV viral genomes they carry, as well as some matching lymphoblastoid cell lines were examined by immunoblotting for the expression of both EBNA1 and EBNA2 antigens. Variation in the molecular weight of EBNA1 indicated that both A and B virus types contained a variety of different virus isolates. EBNA2A was identified in all lines carrying A type viral genomes, but was not observed in any of the lines harboring B type virus. EBNA2B was identified in 4 of 10 Burkitt lymphoma lines carrying EBV type B.

Reactions of sera from patients with rheumatoid arthritis, systemic lupus erythematosus and infectious mononucleosis to Epstein-Barr virus-induced polypeptides.

P3HR-1 and Ramos cells induced with sodium butyrate and 12-O-tetradecanoylphorbol 13-acetate were used in the protein immunoblot technique to identify Epstein-Barr virus (EBV)-specific antibodies present in sera from clinically normal individuals and patients with systemic lupus erythematosus (SLE), rheumatoid arthritis (RA) and infectious mononucleosis (IM). Sixteen EBV-specific polypeptides were detected ranging in mol. wt. from 22,000 (22K) to 140K. Many of the sera contained antibodies to different subsets of these antigens, and a high proportion expressed autoantibodies which reacted with cellular components from an EBV genome-negative cell line. About 50% of the sera from each category reacted with the 44K to 48K and 36K and 38K early antigen (EA) components. A high proportion of the SLE sera (64%) were found to contain anti-EA antibodies, suggesting an association between EBV and SLE. Almost all of the EBV-seropositive sera examined contained antibodies against a 22K late antigen, but none of the sera from IM patients reacted with this polypeptide.

Identification of Epstein-Barr virus-induced polypeptides in P3HR-1 cells by protein immunoblot.

The protein immunoblot technique was used to identify Epstein-Barr virus-specific antigens present in sodium butyrate-induced P3HR-1 cells. Using sera from patients with either nasopharyngeal carcinoma or arthritis, 16 polypeptides were detected ranging in molecular weight from 22K to 140K. Each of the anti-EA-, anti-VCA-positive sera were found to contain antibodies to different subsets of the antigens. A 72K protein was identified which was consistent with the nuclear antigen (EBNA), and culturing cells in the presence of disodium phosphonoacetate allowed identification of 140K and 22K antigens as late viral products. Treatment of cells with sodium butyrate revealed that expression of some antigens increased in parallel with the time of incubation of the cells in butyrate while other antigens either appeared early and then decreased in intensity or were only present after a number of days of butyrate treatment. One of the antigens which decreased with the time cells were treated with butyrate was EBNA.