RESUMO
The lead optimization process in drug discovery campaigns is an arduous endeavour where the input of many medicinal chemists is weighed in order to reach a desired molecular property profile. Building the expertise to successfully drive such projects collaboratively is a very time-consuming process that typically spans many years within a chemist's career. In this work we aim to replicate this process by applying artificial intelligence learning-to-rank techniques on feedback that was obtained from 35 chemists at Novartis over the course of several months. We exemplify the usefulness of the learned proxies in routine tasks such as compound prioritization, motif rationalization, and biased de novo drug design. Annotated response data is provided, and developed models and code made available through a permissive open-source license.
Assuntos
Inteligência Artificial , Química Farmacêutica , Química Farmacêutica/métodos , Intuição , Descoberta de Drogas/métodos , Desenho de Fármacos , Aprendizado de MáquinaRESUMO
BACKGROUND: Subcutaneous implantable cardioverter defibrillator (S-ICD) is a suitable alternative for transvenous ICD (TV-ICD) patients who have undergone transvenous lead extraction (TLE). Limited data are available on the outcome of S-ICD patients implanted after TLE. We assessed the safety, efficacy, and outcome of S-ICD implantation after TLE of TV-ICD. METHODS: The study population consisted of 36 consecutive patients with a median age of 52 (44-66) years who underwent S-ICD implantation after TLE of TV-ICD. RESULTS: Indications for TLE were infection (63.9%) and lead malfunction (36.1%). During a median follow-up of 31 months, 3 patients (8.3%) experienced appropriate therapy and 7 patients (19.4%) experienced complications including inappropriate therapy (n = 4; 11.1%), isolated pocket erosion (n = 2; 5.5%), and ineffective therapy (n = 1; 2.8%). No lead/hardware dysfunction was reported. Premature device explantation occurred in 4 patients (11%). Eight patients (22.2%) died during follow-up, six of them (75%) because of refractory heart failure (HF). There were no S-ICD-related deaths. Predictors of mortality included NYHA class ≥ 2 (HR 5.05; 95% CI 1.00-26.38; p = 0.04), hypertension (HR 22.72; 95% CI 1.05-26.31; p = 0.02), diabetes (HR 10.64; 95% CI 2.05-55.60; p = 0.001) and ischemic heart disease (HR 5.92; 95% CI 1.17-30.30; p = 0.01). CONCLUSION: Our study provides evidences on the use of S-ICD as an alternative after TV-ICD explantation for both infection and lead failure. Mortality of S-ICD patients who underwent TV-ICD explantation does not appear to be correlated with the presence of a prior infection, S-ICD therapy (appropriate or inappropriate), or S-ICD complications but rather to worsening of HF or other comorbidities.
RESUMO
INTRODUCTION: During the coronavirus disease-19 (COVID-19) outbreak in spring 2020, people may have been reluctant to seek medical care fearing infection. We aimed to assess the number, characteristics and in-hospital course of patients admitted for acute cardiovascular diseases during the COVID-19 outbreak. METHODS: We enrolled all consecutive patients admitted urgently for acute myocardial infarction, heart failure or arrhythmias from 1 March to 31 May 2020 (outbreak period) and 2019 (control period). We evaluated the time from symptoms onset to presentation, clinical conditions at admission, length of hospitalization, in-hospital medical procedures and outcome. The combined primary end point included in-hospital death for cardiovascular causes, urgent heart transplant or discharge with a ventricular assist device. RESULTS: A similar number of admissions were observed in 2020 (Nâ=â210) compared with 2019 (Nâ=â207). Baseline characteristics of patients were also similar. In 2020, a significantly higher number of patients presented more than 6âh after symptoms onset (57 versus 38%, Pâ<â0.001) and with signs of heart failure (33 versus 20%, Pâ=â0.018), required urgent surgery (13 versus 5%, Pâ=â0.004) and ventilatory support (26 versus 13%, Pâ<â0.001). Hospitalization duration was longer in 2020 (median 10 versus 8 days, Pâ=â0.03). The primary end point was met by 19 (9.0%) patients in 2020 versus 10 (4.8%) in 2019 (Pâ=â0.09). CONCLUSION: Despite the similar number and types of unplanned admissions for acute cardiac conditions during the 2020 COVID-19 outbreak compared with the same period in 2019, we observed a higher number of patients presenting late after symptoms onset as well as longer and more complicated clinical courses.
Assuntos
Arritmias Cardíacas/epidemiologia , COVID-19/epidemiologia , Insuficiência Cardíaca/epidemiologia , Hospitalização/estatística & dados numéricos , Infarto do Miocárdio com Supradesnível do Segmento ST/epidemiologia , Doença Aguda , Idoso , Idoso de 80 Anos ou mais , Feminino , Hospitais de Ensino , Humanos , Itália/epidemiologia , Masculino , Pessoa de Meia-Idade , PandemiasRESUMO
The ionization state of drugs influences many pharmaceutical properties such as their solubility, permeability, and biological activity. It is therefore important to understand the structure property relationship for the acid-base dissociation constant pKa during the lead optimization process to make better-informed design decisions. Computational approaches, such as implemented in MoKa, can help with this; however, they often predict with too large error especially for proprietary compounds. In this contribution, we look at how retraining helps to greatly improve prediction error. Using a longitudinal study with data measured over 15 years in a drug discovery environment, we assess the impact of model training on prediction accuracy and look at model degradation over time. Using the MoKa software, we will demonstrate that regular retraining is required to address changes in chemical space leading to model degradation over six to nine months.
Assuntos
Fenômenos Químicos , Aprendizado de Máquina , Modelos Teóricos , Reprodutibilidade dos TestesRESUMO
Identification of metabolic biotransformations can significantly affect the drug discovery process. Since bioavailability, activity, toxicity, distribution, and final elimination all depend on metabolic biotransformations, it would be extremely advantageous if this information could be produced early in the discovery phase. Once obtained, this information can help chemists to judge whether a potential candidate should be eliminated from the pipeline or modified to improve chemical stability or safety of new compounds. The use of in silico methods to predict the site of metabolism in phase I cytochrome-mediated reactions is a starting point in any metabolic pathway prediction. This paper presents a new method, specifically designed for chemists, that provides the cytochrome involved and the site of metabolism for any human cytochrome P450 (CYP) mediated reaction acting on new substrates. The methodology can be applied automatically to all the cytochromes for which 3D structure is known and can be used by chemists to detect positions that should be protected in order to avoid metabolic degradation or to check the suitability of a new scaffold or prodrug. The fully automated procedure is also a valuable new tool in early ADME-Tox assays (absorption, distribution, metabolism, and excretion toxicity assays), where drug safety and metabolic profile patterns must be evaluated as soon, and as early, as possible.