From artificial neural networks to predictive modelling, artificial intelligence (AI) is making its mark in the pharmaceutical industry. The ever-increasing cost of development, shorter timelines and rise in demand has seen the industry reach out to AI to help bring drugs to market faster and cheaper. Machine learning (ML) is one branch of AI with exciting applications in drug discovery. 

For daily articles on the latest pharma trends and innovations, as well as interviews with leading experts and in-depth industry White Papers, subscribe to PharmaFeatures.com.

The cost of bringing a drug to market continues to increase at an exponential rate. According to a recent analysis, between 2009 and 2018, US biopharmaceutical companies spent approximately $1 billion bringing each new drug to market. The majority of large expenses occur during the early phases of drug development in drug discovery

Hear from some of the industry leaders including Huijun Wang – who will be providing her expertise in leading a discussion Expediting the Drug Discovery Process Through Chemistry and Biology: Leveraging AI in Hit Finding and Lead Optimisation. To discuss these innovations and more with other leading experts in an informal setting, sign up to Proventa’s Medicinal Chemistry and Biology Strategy Meetings, held online on 29 June 2021.  

Drug discovery 

Human dose production is a particularly challenging area, in which conventional in vitro modelling faces a dilemma with poor translatability. This often results in early termination of clinical trials when drugs do not demonstrate the same pharmacokinetics within humans as predicted preclinically. 

It appears that ML, a branch of AI, is leading the way in the latest innovations. A recent example was a 2021 study which used ML attempts for predicting human subcutaneous bioavailability of monoclonal antibodies. The measured bioavailability of the monoclonal antibodies ranged from 35% to 90%. The decision tree-based method, a form of ML, proved to best predict bioavailability. 

Since all of the ML approaches used theoretical calculations and predictions for input, it was suggested from the study that these models may be most useful for early-stage activities like molecule formational design. 

A form of AI known as natural language processing (NLP), can be used to optimise the process of target identification. NLP extracts “meaning from human language to make decisions based on the information”. This can be used to scan vast numbers of publications and genetic databases to search for gene-disease associations and identify new targets. The AI-based algorithms can perform tasks such as this with greater accuracy and speed in comparison with human intelligence. 

The importance of prioritising the most potent compounds for a relevant therapeutic target is emphasised in a 2018 study investigating machine learning for predicting drug-target interactions.

In the publication, it is emphasised that hit identification “is the first step towards new drug development. Identifying unexpected off-targets can open the possibility of drug repurposing or can lead to insights for predicting and explaining observed side-effects.”

Machine learning on DNA-encoded libraries

DNA-encoded libraries (DELs) have been increasingly explored in recent years to enhance hit identification in drug discovery. DELs represent a modern and versatile tool used to better identify a greater range of novel biological compounds. These libraries are capable of screening drug targets with an extensive number of compounds with great efficiency.

Unfortunately, in order to analyse vast amounts of data, DELs are still dependent on bioinformatics operated by humans. The result of this “limits the scale of molecules considered, introduces bias, and makes it difficult to fully utilise the subtle patterns in the DEL selections”.

Utilising ML allows the identification of important features and obvious patterns from a small dataset and uses the information to create projections for larger datasets. In a recent study, two types of ML models were trained on the DEL selection data to classify compounds: random forest and graph convolutional neural network (GCNN). 

The random forest is an algorithm that creates a predictive model comprising a large number of individual ‘decision trees’ which operate as a whole group. Each tree in the forest produces a class prediction and the class with the most votes becomes the model’s prediction. 

These methods have already demonstrated success in a study which reported that ML models verified hits up to 29% at one micromolar. The ability to identify target molecules on a micro scale is critical for creating a larger hit pool.

GCNN is a form of ML known as deep learning. One of the main benefits for the application of GCNN for DEL is that “deep learning methods automatically extract important features from a dataset whereas manually generated features are necessary for conventional machine learning algorithms”. Therefore, GCNN is more likely to identify potential hits faster and with greater accuracy than DEL alone. 

Iterative screening

High throughput screening (HTS) is the most popular approach for screening large libraries of compounds against a target of interest. Unfortunately, the sheer size of these libraries results in a high screening cost to run. In addition, the low hit rate of HTS, typically less than 1% in most assays, requires large compound libraries to generate a sufficient number of hits for drug development programs to progress.  

Iterative screening is a process in which drug screening is performed in batches – each batch is filled by using ML to select the most promising compounds from the library based on the previous results. Iterative screening has been shown to enhance the efficiency of HTS, as it allows for a smaller part of the library to be screened at a time, while still identifying a large portion of the active compounds.

Previously, an iterative approach to HTS was considered impractical due to the high labour costs, however “advances in screening automation have made custom selection of compounds more broadly feasible”.

A recent study investigated the iterative approach to HTS, and found that “the hit rate in the iterative screening was just greater than twice that of normal (random) screening, recovering a median of 78% of the active compounds when 35% of the library had been screened”. 

It is worth noting that there are a number of potential practical challenges that can arise with iterative drug screening. While the iterative approach to screening increases the rate of hit identification, the overall process can be “resource intensive and the interim analysis of screening data will potentially require more time for quality control and data management”. These are considerations that will need to be taken into account when weighing up the value of hit identification and whether these challenges can be overcome in the future. 

The results from this study however, remain positive and demonstrate how ML approaches like iterative screening show potential to optimise drug discovery. 

To discuss these topics further with sector experts, and to ensure you remain up-to-date on the latest in clinical development, sign up for Proventa International’s Medicinal Chemistry and Biology Strategy Meeting, set for 29 June 2021.

Charlotte Di Salvo, Lead Medical Writer
PharmaFeatures

For more articles covering the pharmaceutical industry, clinical research and academia, visit our content site PharmaFeatures.

The post Utilising Machine Learning in Drug Discovery: Opportunities and Challenges appeared first on Proventa International.

Computational modelling is becoming increasingly popular for data analysis in life sciences. Vast areas of therapeutic research are taking advantage of machine learning (ML) approaches for disease predictions and pathology. Cancer image analysis and diabetes case prediction are a few of the latest innovations.

For daily articles on the latest pharma trends and innovations, as well as interviews with leading experts and in-depth industry White Papers, subscribe to PharmaFeatures.com.

Hear from some of the industry leaders including Steven Zimmer who will be providing his expertise in leading a discussion on how multi-targeted drugs could be the solution to low drug discovery productivity. To discuss these innovations and more with other leading experts in an informal setting, sign up to Proventa’s Bioinformatics Strategy Meetings, held online on 1 July 2021.

ML is a branch of artificial intelligence showing exciting applications across drug development. With each exposure to new data, an ML machine-learning algorithm grows increasingly better at recognising patterns over time. There are two main techniques used to apply ML: supervised and unsupervised learning. Unsupervised learning is a type of algorithm that learns patterns from data without tags (annotations). Supervised learning algorithms, on the other hand, are formed from labelled training data which consists of a set of training examples. In others words, supervised learning relies on human intervention to label data in order to train the model to search for a specific component – cancer image analysis for example. Unsupervised learning on the other hand, analyses vast amounts of data which has not been labelled in order to identify associations or trends.

Supervised learning methods have been used to predict future values of data categories or continuous variables. Unsupervised learning is primarily used for exploratory purposes in the development of models to enable data clustering in a format not specified by the user. This particular technique helps to identify hidden patterns within input data, whereas supervised learning methods predict future outputs based on a trained model of known input and output data. 

Diabetes research 

In March 2020, an abstract supplement was published detailing the “development of a machine-learning method for predicting new onset of diabetes mellitus (DM)”. While data predictions are not a novel concept within diabetes research, they typically only apply for those predisposed to health conditions, not healthy individuals. 

Within this study, an ML-based prediction model was used to identify DM signatures prior to onset. Signatures for DM could be biomarkers, for example, or blood-based factors like serum proteins.

These signatures would be identified through the data analysis of nationwide health records of patients from 2008-2018 via the ML prediction-based model. The model utilised a type of ML known as gradient-boosting decision trees. A gradient-boosting decision tree (GBDT) model is typically a prediction-based form of AI used to calculate the likelihood of interactions.

The study identified a total of 4,696 new diabetes patients (7.2%) from datasets. Their ML model predicted the future incidence of diabetes with an overall accuracy of 94.9%.

It is worth noting however that the algorithm for GBDT was originally developed for static data, i.e. fixed size of data. However, in the context of diabetes research, data would be constantly changing with incoming patient medical records. Therefore, it could potentially be a time-consuming and impractical process to run GBDT every time incoming data arrives. In terms of solutions, one article has suggested that GBDT needs to be adapted to “an incremental learning setting, where new samples are continuously arriving in batches”.

Diabetes mellitus is a chronic disease and increases the risk of developing diseases such as cancer and atrial fibrillation, which can be fatal. Hence, predicting diabetes in the population could prevent potential cases through medication or diet control. In the long-term, this would reduce the likelihood of said patients developing serious diseases as a result of diabetes, which would theoretically reduce the pressure on healthcare systems around the globe. 

Immunology

Artificial intelligence has been used as an important tool within immunology to answer highly complex questions. More recently, a study published August 2020 demonstrated that deep learning neural networks can be used to differentiate between immune cells. They specifically used a convolutional neural network (CNN), a form of deep learning. A CNN is a deep learning model inspired by the “animal visual cortex in structure and designed to automatically and adaptively learn spatial hierarchies of features, from low- to high-level patterns”.

The study demonstrated how this ML approach could learn to predict the patterns of chromatin opening across 81 stem and differentiated cells across the immune system, solely from the DNA sequence of regulatory regions. 

Chromatin is the material which constitutes a chromosome composed of DNA and protein. 

Open chromatin regions reflect quite closely gene expression in the corresponding cells, hence why these areas are a target for cell identification in the immune system. 

This deep learning approach has shown to be an important tool for immunology researchers, revealing modalities and complex patterns of immune transcriptional regulators that arise directly from the DNA sequence. Immune transcriptional regulators play a critical role in the maintenance of the immune system. These factors primarily control gene expression for various immune cells, thus have been implicated in autoimmune disorders when the immune system malfunctions. 

This was raised in a 2018 study which emphasised how dysregulated (gene) expression has been correlated to immune cell dysfunction in autoimmunity and lymphomagenesis. Therefore ML approaches like those used in the aforementioned study may help researchers to understand the complex mechanism behind cellular phenotype in the immune system. And, potentially, contribute to therapeutic developments for many immunological disorders.

Oncology 

In addition to immunology, deep learning approaches in oncology are becoming increasingly popular across basic and clinical cancer research.

Deep learning approaches have brought significant advancements to cancer image analysis. Early-stage cancer is often difficult to detect, especially so with conventional technology and human error, thus ML approaches like convolutional neural networks could potentially analyse images with greater speed and accuracy.

There are a number of challenges, however, with the deep learning approach to image analysis. Firstly, differences in colour tone on pathology slides may occur across different institutions due to the type of staining and sample preparation protocols: i.e. it presents an issue if one research lab uses colour x to stain their samples to highlight cancerous regions but the ML model has been trained from images of samples stained with Y, it could be difficult to accurately detect cancer as it is not the same staining. Therefore, it is necessary to “standardize color tones in digital slides for the development of accurate AI algorithms”.

Secondly, the limited number of medical images available for network training is a problem. Data augmentation is one strategy that has been developed by researchers to overcome this issue. Data augmentation i.e. where images are randomly cropped, tilted, inverted or flipped to increase their number, is one effective strategy for dealing with the small size of the training set. 

A 2017 study successfully trained a CNN to classify skin cancer with a level of competence comparable to dermatologists. Using only pixels and disease labels as inputs, they classified skin lesions via a single CNN. The 1.4 million pre-training and training images in this study overcame photographic variability like zoom and lighting.
This is a huge step forward for cancer imaging. The development of an accurate ML model for image analysis could support medical practitioners and patients to “proactively track skin lesions and detect cancer earlier”. Early detection significantly impacts cancer prognosis for many patients and MLapproaches like this could save many lives.

To discuss these topics further with sector experts, and to ensure you remain up-to-date on the latest in clinical development, sign up for Proventa International’s Bioinformatics Strategy Meeting, set for 1 July 2021.

Charlotte Di Salvo, Junior Medical Writer
Proventa International

The post Machine Learning Data Analysis: Applications Across Immunology, Oncology and Diabetes appeared first on Proventa International.

Precision medicine is a rapidly growing therapeutic area within oncology. Technological innovations have contributed to the refinement of predictive and diagnostic techniques used to tailor therapy for cancer patients. The goal is to not only develop more precise treatments, but to also predict the optimum choice of therapy to maximise the chance of progression free survival.

For daily articles on the latest pharma trends and innovations, as well as interviews with leading experts and in-depth industry White Papers, subscribe to PharmaFeatures.com.

Deep learning for diagnostics 

Deep learning is a specialised area of ML that attempts to model abstraction from large-scale data using multi-layered deep neural networks (DNNs). Abstraction refers to the process of filtering out irrelevant data in order to focus on the desired information. Cancer diagnosis relies upon accurate visual pattern recognition in images and large data sets in order to identify features that cause concern. While oncology physicians are highly specialised and trained in their field, there is only so much data processing that the human brain can handle. Machine learning on the other hand can be programmed to identify abnormalities in vast data sets with greater accuracy land speed than the average human. 

Deep learning has been used for image classification in cancer diagnosis. Such image tasks require a form of deep learning known as a convolutional neural network (CNN). This artificial neural network relies on many layers composed of connected artificial neurons that perform mathematical operations on input data. One of the main reasons why CNN works well for image classification is the ability of the network to mimic the natural visual processing of the human brain “which enables the interpretation of dense information such as the relationship of nearby pixels and objects”. 

During training, labeled image data is inputted into the artificial neural network and undergoes two processes, known as filtering and sub-sampling. These processes enable the network to learn the image features. Labelled image data may comprise images of skin lesions with cancerous features identified. Once a model is trained, it is validated in an independent rest in order to evaluate its final performance.

In 2020, a study published in Nature Research used deep learned tissue “fingerprints” to classify breast cancers. One of the challenges of using deep learning for histopathology was noted in the study as the lack of large, well-annotated data sets required for the algorithms to learn statistical significance. After training the algorithm, they used “the features the network learned, called ‘fingerprints,’ to predict ER, PR, and Her2 status in two datasets”. While the dataset for the study was small, the fact that the fingerprints determined different growth factor status from whole slide images of breast cancer. This supported the potential of using ML in cancer diagnosis. 

Next-generation sequencing 

Next-generation sequencing (NGS) is a DNA-sequencing technique used for high-throughput tumour profiling. With the ability to sequence the entire human genome in a day, NGS offers a significant advantage over traditional genomic sequencing like the Sanger sequencing. According to a recent review, NGS comprises of the following steps:

• Each DNA fragment to be sequenced is bound to a structure called array, followed by the enzyme DNA polymerase which adds labeled nucleotides sequentially. 

• A high-resolution camera captures the signal from each nucleotide as it becomes integrated and notes the spatial coordinates and time. 

• The sequence at each spot can then be inferred by a computer program to generate a contiguous DNA sequence, referred to as a read.

NGS is an important part of genetic sequencing within a diagnostic technique known as liquid biopsy. Liquid biopsy is a non-invasive and real-time monitoring of disease development, which can be applied to all stages within cancer diagnosis and treatment. Liquid biopsy measures the presence of circulating tumour DNA (ctDNA). cTDNA is a type of cancer biomarker used to detect the disease as well as monitor the progress throughout treatment. 

NGS is used in conjunction with liquid biopsies to provide a tumour-specific molecular profile of the cancer. Initially PCR-based methods were used to sequence ctDNA due to their sensitivity and low cost. However, these methods can only screen for known variants, and the input and speed are limited. NGS has shown to provide high throughput and the ability to screen unknown genetic variants. Thanks to NGS technology, the sequencing of ctDNA can be performed at a much higher sensitivity than tissue biopsies and support patients in targeted therapy relative to their tumour profile. 

Microsatellite instability testing

The DNA mismatch repair system (MMR) is a highly conserved repair mechanism for cellular function. However, when the MMR system fails to work properly, it can cause microsatellites. These are regions of repeated DNA that change in length, showing instability.  Lynch syndrome is a common hereditary disease, characterised by mutations in MMR genes. This syndrome is associated with many cancer types, especially so with colon cancer and endometrial cancer.

MSI testing is used to analyse the length of specific DNA microsatellites within a tumour sample in order to measure the level of instability.  

Clinical studies in the last five years suggest that MSI is a predictive biomarker for immunotherapy, seeking further attention for its application in precision medicine. According to a 2019 review, several clinical trials have demonstrated that mismatch repair deficiency or microsatellite instability-high is significantly associated with long-term immunotherapy-related responses and better prognosis in colorectal and noncolorectal malignancies treated with immune checkpoint inhibitors. 

At the time of publication (2019), the drug pembrolizumab has been approved for MMR deficiency/microsatellite instability-high refractory tumours, and nivolumab approved for colorectal cancer patients with MMR deficiency / microsatellite instability-high. The fact that the same biomarker has been used to support immunotherapy for different tumour types is a first in cancer research. This represents a significant step forward in precision oncology, highlighting a promising opportunity to improve the efficacy of immunotherapy.

The three innovations described in this article are a few of the latest developments in precision oncology. As research continues, the hope is that cancer therapy will become more targeted so treatment is more effective and patients receive the best chance of survival.

To discuss these topics further with sector experts, and to ensure you remain up-to-date on the latest in clinical development, sign up for Proventa International’s Oncology Strategy Meeting, set for 17 June 2021.

Charlotte Di Salvo, Junior Medical Writer
Proventa International

The post The Latest Developments in Precision Oncology appeared first on Proventa International.

Risk-based monitoring (RBM) is a useful tool in maximising the quality of clinical trials. Enhanced site communication, greater patient safety and lower costs are a few of the benefits of RBM. The innovations of AI in clinical research have application in RBM, optimising further the design and conduct of clinical trials. 

For daily articles on the latest pharma trends and innovations, as well as interviews with leading experts and in-depth industry White Papers, subscribe to PharmaFeatures.com.

RBM allows sponsors to identify and address potential issues that could comprise the quality of a clinical trial. Identifying, assessing, monitoring and minimising any potential risks is of paramount importance to prevent delay or termination of clinical trials. 

There are three key parts of RBM:

• Identify critical data and processes

• Risk assessment

• Monitoring plan

According to FDA guidance, the risk assessment serves to “identify and understand the nature, sources, likelihood of detection, and potential causes of risks that could affect the collection of critical data or performance of critical processes”. The more information about high-risk areas for potential issues or critical data points, the easier a monitoring plan can be developed. In the FDA document, there are a number of points raised when determining the type of monitoring required. The complexity of study design; study endpoints; clinical complexity of study population; and geographic location of sites are four of the nine factors that need to be considered for RBM. 

In addition to the obvious identification of a risk and its origin, re-education of the site and amendment to recruitment plans are other ways in which a risk assessment can be performed. 

In order to identify the critical data points at the start of the clinical trial, it is important to identify the variables that need to be measured to answer the original scientific question of the study. Critical data points in clinical trials can include anything from the target parameters, to compliance for subject criteria to serious adverse events. 

A white paper by IQVIA highlights a number of statistics in support of RBM for clinical trial efficiency, including “developed RBM solutions can bring as much as 25% cost reduction over traditional trial execution approaches”.

Application of AI in RBM

Predictive analytics

Predictive analytics is one of the more recent innovations in RBM that incorporates AI technology. It is the practice of extracting information from existing data in order to determine patterns and predict future outcomes and trends. In terms of risk management, it can be used to predict future scenarios with an acceptable level of reliability, in addition to ‘what-if’ scenarios which can be used to develop a risk assessment. 

Machine learning (ML) is an important part of clinical predictive analytics. Due to the accessibility of electronic health records (EHR), the volume of data in clinical research has increased substantially. Therefore, computational software has become widely used as a more cost- and time-efficient method of analysing vast amounts of data.

In 2018, a retrospective, single-site study investigated the use of ML to identify high-risk surgical patients using automatically curated electronic health record data (Pythia). The first aim of the study was to develop a ML model that could analyse high-volume, high-quality data (to monitor care and patient outcome). The second was to promote the development of ML models that can be used to interpret clinical data, and support clinicians in identifying potential high-risk patients. 

Random forest and extreme gradient boosted decision trees were a few of the ML methods used to predict the likelihood of post-surgical complications. In addition, a method known as Lasso penalized logistic regression was used due to the ability of the algorithms to quantify the importance of variables in a dataset. The study emphasised that “by providing model users with additional information about predictor weights, clinicians can glean insights into potential patient risk mitigation strategies”. 

The positive results demonstrated the efficacy of utilising ML for predictive analytics. The 42 models used demonstrated “strong predictive performance”, with the random forests showing the greater. The prediction of high-risk patients in this study reinforces the translatability to RBM in clinical research. Using predictive analytics could help forecast potential issues which can be addressed and alter risk assessments accordingly. 

AI Data Security

Thanks to the introduction of digital innovations like telehealth, the amount of data within clinical research is increasing substantially. Furthermore, the pandemic has seen a rapid rise in the number of decentralised clinical trials globally. This new era of virtual clinical research has brought a number of advantages, streamlining clinical trials with cloud-based software. However, with confidential data like EHRs being shared across multiple platforms, data security has become a priority within RBM. 

Blockchain is an example of technology which has converged with AI and is becoming increasingly popular for data security within clinical trials. The traditional centralised server models are vulnerable to the single-point attack limitations and malicious insider attacks. However, blockchain technology spreads data across a large network of databases in replica copies, rather than a single store. The benefit of this is that stored data is less vulnerable to hacking or infringement. Furthermore, the incorporated verification ensures data is protected from unauthorised access. In terms of the convergence with AI, blockchain can provide privacy and trustworthiness, and AI utilises machine learning algorithms on the blockchain to achieve security, scalability and effectiveness. 

It is worth noting however, blockchain technology still has some limitations. Expensive software, large storage and high bandwidth are a few of the challenges with blockchain which may be not suitable for the smaller pharmaceutical companies and contract research organisations. 

With the help of AI, RBM will continue to evolve at the same pace as the digitalisation of clinical research. The value of predictive analytics has reinforced the importance of developed models to enhance RBM for clinical trials, reducing the time and cost spent addressing issues that could have been predicted earlier.

To discuss these topics further with sector experts, and to ensure you remain up-to-date on the latest in clinical development, sign up for Proventa International’s Clinical Operations Strategy Meeting, set for 15 June 2021.

Charlotte Di Salvo, Junior Medical Writer
Proventa International

The post Innovations in AI Risk-Based Monitoring in Clinical Research appeared first on Proventa International.

Dose prediction is a critical part of drug discovery which determines the safety and efficacy of drugs within patients. Dose response predictions are made with preclinical models which quantify pharmacokinetic (PK) behaviour via a number of parameters. Despite the continuous use of in vitro and in vivo models, it appears that automated, machine learning approaches are leading the way in the latest innovations of human dose prediction. 

For daily articles on the latest pharma trends and innovations, as well as interviews with leading experts and in-depth industry White Papers, subscribe to PharmaFeatures.com.

Introduction

There are two main factors linked to dose prediction: PK and drug-drug interactions. Pharmacokinetics is the behaviour of a drug within a system, and includes bioavailability, distribution, metabolism and excretion. Drug-drug interactions (DDI) can occur when multiple drugs are co-administered to a patient. DDI is an important part of drug development, as it can lead to altered systemic exposure, causing variable drug responses. Typically DDIs occur due to inhibition of the metabolism for one drug by the other; it leads to a rise in plasma concentration of the drug whose metabolism is inhibited.  

Preclinical data helps to inform of the likelihood of achieving the desired therapeutic dose in early clinical development. From here, the prediction methods are assessed and further refined from the comparison of clinical data with the original predictions. Due to continuous developments in predictive methodologies and data analysis, the probability of successful PK predictions has increased over the past few years. According to a 2020 study, 83% of AstraZeneca drug development projects progress in the clinic with no PK issues. 

The higher the probability of successful predictive models, the faster the progression from drug discovery to market. Poor prediction of drug behaviour in clinical trials can significantly hinder drug development, either due to poor efficacy or toxicity issues. In terms of safety testing, drugs are often tested in concentration-response assays. The results are quantified by an XC50 value, which is a measure of the potency of the compound, and which lies between the lowest and highest concentrations tested. 

Potency is a key component of predicting PK and is a measure of drug activity relative to the amount of drug required to produce a desired effect. Another important assessment is the maximum concentration of a compound when it reaches the target, known as Cmax. This refers to the clinical dose, and compounds given at higher doses tend to have a greater risk of toxic side effects, making Cmax one of the best predictors of clinical toxicity. 

Efficacy of current models

Through a combination of in silico, in vitro and in vivo assays, there are four PK parameters a drug must pass in order to inform about its pharmacokinetic activity for dose prediction.

Prediction of volume distribution (Vd) is a measure of the relative affinity of a drug for tissues and plasma. Vd is reasonably predicted from in silico and in vitro modelling. Using preclinical in vivo PK data, typically animal models, Vd is the most predictable of the PK parameters. 

The prediction of drug clearance (CL) uses a combination of animal data and human in vitro modelling. The primary CL routes in humans and preclinical species are hepatic metabolism and renal and biliary elimination, and methods are required to predict these parallel elimination processes in humans. Identifying the elimination route and rate in preclinical species relative to human CL makes this one of the most challenging parameters in drug discovery. 

Allometry is a method used to predict drug CL and is still widely used within the industry. According to a 2016 study, allometric scaling is where the “exchange of drug dose is based on normalization of dose to body surface area”. In other words, it is used to scale physiological rates relative to the growth and size of one (animal) species relative to another. 

Prediction of bioavailability is another key parameter and refers to the extent and rate at which a drug accesses the target site. Poor bioavailability comprises the efficacy of a drug if the desired therapeutic dose is not correct once the drug reaches its target. In addition, low bioavailability can create issues in obtaining appropriate exposures in pharmacodynamic (PD), efficacy, and safety experiments for a candidate drug. 

Finally, the prediction of pulmonary absorption is an important part of optimising drugs delivered directly to the lungs. The two main strategies for optimising drug retention in the lungs after inhalation involves maintaining a drug reservoir within the airways and a slow rate of dissolution. 

Innovations in predicting human dose

The industry has seen continuous developments in in vitro/vivo/silico models in order to optimise human dose production. However it appears that computational biology and AI are leading the way in the latest innovations. 

In immunology, a 2021 study used machine learning (ML) attempts for predicting human subcutaneous bioavailability of monoclonal antibodies. The aim of this study was to develop a ML approach to determining whether the bioavailability of monoclonal antibodies was >70%. The measured bioavailability of the monoclonal antibodies ranged from 35% to 90%. The team used multiple ML methods including random forest and decision trees. These tree-based methods proved to best predict bioavailability. 

Since all of the ML approaches used theoretical calculations and predictions for input, it was suggested from the study that these models may be most useful for early-stage activities like molecule formational design. 

Oncology is another therapeutic area which has utilised automated methods for dose prediction. Intensity modulated radiation therapy is a deep learning approach which delivers specific doses to target tissue while sparing adjacent organs at risk. This is an important part of cancer treatment, by helping to reduce radiation-related toxicities. 

In 2020, a study investigated fully automated dose prediction using generative adversarial networks in prostate cancer patients. The preliminary study focused on CTs of prostate cancer patients using a generative adversarial network framework, the choice of which was suggested to be due to the “variations of the targets in the prostate cancer patients [being] relatively small”. 

The future for ML methods is suggested to optimise treatment strategy before radiation, as radiotherapy can cause complications with organ placement and re-irradiation.

To discuss these topics further with sector experts, and to ensure you remain up-to-date on the latest in clinical development, sign up for Proventa International’s Medicinal Chemistry and Biology Strategy Meeting, set for 29 June 2021.

Charlotte Di Salvo, Junior Medical Writer
Proventa International

The post How Effective is the Automation and Validating of Human Dose Predictive Models? appeared first on Proventa International.

The modernisation of clinical research is an important part of expediting the drug approval process, and this in turn is dependent on the success of clinical trials with a sufficient number of appropriate participants. A number of recent innovations promise to bring about this modernisation: for example, using artificial intelligence (AI) software to research medical databases for eligible participants reduces the time scale for recruitment. Telemedicine will encourage more patients to take part in clinical trials from the comfort of their own home. But as considerable a solution as they are, these new technologies do have their own challenges to face.

For daily articles on the latest pharma trends and innovations, as well as interviews with leading experts and in-depth industry White Papers, subscribe to PharmaFeatures.com.

Telemedicine and virtual clinical trials

Telemedicine is the use of remote technology, typically for clinical consultations and the delivery of healthcare. It enables communication between patients and healthcare providers outside the clinical environment. One example of telemedicine is the recording of vital signs during the sleep cycle using actigraphy bracelets, for example Fitbit. 

Telemedicine has facilitated clinical researchers in capturing and analysing data remotely. This method has been desirable for both patients and clinical organisations. Firstly, limited travelling to clinics is more likely to increase patient retention in clinical trials, allowing them to continue day-to-day life. Secondly, the data is more likely to be representative of real-word evidence, in comparison to clinics which do not always show a patient’s true behaviour in an unknown, clinical environment. 

Telemedicine has shown obvious potential in clinical trials for rare diseases. Rare diseases typically have small populations across the entire globe, hence a central site for clinical trials may not be accessible. In March 2017, the European Reference Networks (ERNs) were launched “as virtual networks enabling healthcare providers across Europe to access and share expertise for the care of patients with complex or rare disorders.” 

According to a 2017 publication, the ERN is composed of “at least ten healthcare providers from at least eight different Member States”. In addition to addressing the small number of patients, it enables the liaison of medical experts across the world. This offers an advantage for rare diseases, a therapeutic area which typically lacks knowledge of disease pathology and therapeutic options. 

The increasing number of virtual trials that emerged during the pandemic utilised telemedicine to continue clinical trials, reducing patient-clinician contact in COVID-19 risk assessments. In comparison to traditional clinical trials, patient recruitment, consent and data collection all occur virtually. Face-to-face clinical appointments are often eliminated altogether, along with physical sites. 

The development of virtual trials, also known as decentralised clinical trials, demonstrates the industry evolving rapidly to “deliver approaches that reduce patient burden, increase patient engagement, and promote trial continuity”. The success of virtual clinical trials throughout the pandemic demonstrated the benefits of a patient-centric approach to clinical research. 

One of the challenges to overcome with the digitalisation of clinical trials is the skepticism surrounding data security. The majority of clinical data is highly confidential, with data sources like electronic health records containing a patient’s entire medical history. 

Blockchain is the latest advancement in data security within clinical research. Blockchain technology is a shared system for recording transactions, tracking assets and building trust in a network. The benefit of this is that stored data is less vulnerable to hacking or infringement. Furthermore, the system involves verification steps which ensures the data is protected against unauthorised intervention. 

Artificial intelligence in patient recruitment

Patient recruitment is one of the many areas in clinical trial design streamlined by AI. Natural language processing (NLP) is a branch of AI that enables computers to analyse the written and spoken word. In the context of medicine, it has been used to “allow algorithms to search doctors’ notes and pathology reports for people who would be eligible to participate in a given clinical trial.” In a report by Nature, it is suggested that refinement of NLP could be used in clinical trials to search patient databases for eligible participants. The inclusion and exclusion criteria of clinical trials is typically written in plain text, so shouldn’t require complex algorithms like those required to analyse doctors’ notes.   

In addition to patient recruitment, AI models can also be used to enhance cohort selection. Electronic phenotyping is a well-established discipline which focuses on “reducing population heterogeneity, namely the process of identifying patients with specific characteristics of interest.” Using electronic medical records, “individuals with an explicit observable trait from large quantities of imperfect clinical patient data” can be identified, also known as phenotyping. 

This method is primarily used to reduce patient population heterogeneity rather than enhancing the quality of prognoses. Despite this, phenotyping with EMR data presents a number of challenges including “variation in the accuracy of codes, as well as the high level of manual input required to identify features for the algorithm and to obtain gold standard labels”.

There are a number of other challenges that exist at present. Firstly, a challenge for the pharmaceutical industry is the lack of personnel to operate AI/ML-based platforms. Furthermore, there is often skepticism about the quality of data generated by AI. Small organisations are often limited in their budget so cannot afford to invest in AI/ML technology. 

Cloud-based software

An optimal clinical trial management system (CTMS) is an important project management tool that supports pharmaceutical and clinical operations. A CTMS maintains contract and payment systems, document management, study milestones and contact management for sites and teams. Legacy CTMS’ are essentially outdated systems within clinical trial management. Recent events highlight the benefits of virtual clinical trials, but current CTMS’ will struggle to manage increasing amounts of data from many different platforms. 

Because of this, the industry is moving towards Cloud-based platforms. Cloud-based applications offer a unified clinical platform to collate documents and data, direct sharing of information with external partners and being able to configure a CTMS system to any type of clinical trial.

Innovations in AI and modern technology are going a long way to address some of the issues of traditional clinical research, including poor patient recruitment. While these innovations do come with their own hurdles to overcome, it is without doubt that in the future they will be refined to improve further.

To discuss these topics further with sector experts, and to ensure you remain up-to-date on the latest in clinical development, sign up for Proventa International’s Clinical Operations Strategy Meeting, set for 15 June 2021.

Charlotte Di Salvo, Junior Medical Writer
Proventa International

The post The Impact of Modern Technology on Clinical Operations appeared first on Proventa International.

In the last decade, innovations in data technology have greatly contributed to the quality and transparency of clinical research. System security, clinical trial design, and data analysis are some of the areas which have seen significant modernisation to keep up with clinical demand. The following examples are some of the data innovations emerging within clinical research with a possible transition away from traditional methodology. 

For daily articles on the latest pharma trends and innovations, as well as interviews with leading experts and in-depth industry White Papers, subscribe to PharmaFeatures.com.

Blockchain

Blockchain technology is a shared system for recording transactions, tracking assets and building trust in a network. Data is spread across a large network of databases in replica copies, rather than a single store. The benefit of this is that stored data is less vulnerable to hacking or infringement. Furthermore, the system involves verification steps which ensures the data is protected against unauthorised intervention. 

These properties have made blockchaining a desirable asset for clinical research. The storage and processing of sensitive data such as patient medical records and clinical trial data is of paramount importance. As a secure, distributed datastore, blockchain has the potential to provide the data transparency that clinical research needs. A 2019 Frontier article reinforces this point describing how such an approach could help improve the transparency and trustworthiness of clinical trials and benefit the whole clinical research ecosystem”. There are three fundamental properties of blockchain that make it desirable for clinical studies: time-stamping, time-ordering and smart contracts. 

Within the decentralised store of blockchain are ordered records called blocks. Each block is time-stamped and can be updated by a majority of users. The information cannot be erased and the datastore is not ruled by any trusted third party. The benefit of time-stamping emphasises the integrity of data storage which is consistent and secure. In terms of time-ordering, “event consistency allows for checking the integrity of all time-ordered events”. 

Blockchain creates a datastore which is incorruptible and traceable.This prevents “a posteriori reconstruction and allows for securely automating the clinical trial”. These are also known as Smart Contracts. These three core functionalities ensure clinical data is controlled and secure with shareable parameters whether it be for patients or clinical trial stakeholders.

Data mining

According to IBM, data mining is the process of “uncovering patterns and other valuable information from large data sets”. This is also known as knowledge discovery in data (KDD). Data mining techniques can be used to either describe a target dataset (descriptive) or predict outcomes via machine learning (predictive). 

There are four main steps in data mining: “setting objectives, data gathering and preparation, applying data mining algorithms, and evaluating results”. All four steps in sequence enable the mining of data and identification of patterns without prior research and design. 

Data mining is an important tool in clinical research, identifying trends and patterns within stored medical records and follow-up data. The main benefits of data mining in clinical research include searching for potential relationships from medical data to gain effective knowledge of diagnosis and treatment of patients in addition to increasing the accuracy of disease prediction. 

Association analysis is an example of a descriptive data mining technique used to search for frequent patterns or correlations between data within large datasets. It is worth noting that data mining is not expected to replace traditional research methods, but rather as an extension of statistical analysis. 

Outlier detection models signify novelty, anomaly, noise, variation or could be categorised under mining exceptions. In the context of clinical research, this data mining technique could be used to identify human error or negligence in clinical data entry. This is an important part of reducing inaccurate data that could otherwise lead to a misconception of results. In addition to detecting incorrect data, the presence of outliers could lead to novel insights in clinical knowledge discovery.

Disruptive technologies

Beyond these traditional clinical research methods are so-called ‘disruptive innovations’, of which there are many examples seen recently. One of the most obvious is the emerging popularity of telemedicine. Digital health tools like Fitbit watches allow the collection of huge volumes of data that can be accessed remotely by clinicians. These devices are user friendly, allowing subjects to perform as they would day-to-day. The benefit of this enables the collection of raw, real-world data in comparison to clinic visits which may not reflect true behaviour or physiology. 

Mobile applications are frequently being used to report patient outcomes. Patient engagement apps also enable participants to become more involved in clinical trials, with access to documentation which explains complex clinical and scientific terminology. Informed consent can also be collected electronically. These innovations provide patients with the accessibility and transparency they require to feel comfortable throughout clinical trials, potentially improving overall patient retention. Digital health technologies may give trial participants a choice of participating from the convenience of the home rather than traveling to a trial site, which can increase participant engagement and retention.

Decentralised clinical trials (DCTs) offer a more patient-centric approach to clinical research.

Digital health is a key components within decentralised clinical trials which collect data safely and efficiently from the comfort of a patient’s home. The number of virtual clinical trials that emerged during the pandemic allowed clinical research to continue with limited site visits and patient recruitment. In terms of clinical operations, DCTs can reduce the time and financial burden of managing patients at sites across the globe.

In unprecedented circumstances, digital innovations in clinical research enabled clinical trials to continue running in extenuating circumstances. However, the success of virtual clinical trials also demonstrated the benefits of a patient-centric approach to clinical research. Disruptive technologies, data mining and blockchain technologies are a few examples of the digital innovations that have streamlined clinical trials, transitioning from traditional to modernised operations.

To discuss these topics further with sector experts, and to ensure you remain up-to-date on the latest in clinical development, sign up for Proventa International’s Clinical Operations Strategy Meeting, set for 15 June 2021.

Charlotte Di Salvo, Junior Medical Writer
Proventa International

The post Blockchain, Data mining, and Disruptive Technologies: Digital Innovations in Clinical Research appeared first on Proventa International.

Significant recent investment in computational technology has seen a number of new innovations arise in drug discovery – perhaps most notable machine learning (ML). By 2022, it is expected that AI technology will contribute $2.199 billion to pharma’s revenue, with popularity growing across the pharmaceutical industry. Target identification, validation and drug discovery are some of the areas in drug development in which machine learning has shown its potential. 

For daily articles on the latest pharma trends and innovations, as well as interviews with leading experts and in-depth industry White Papers, subscribe to PharmaFeatures.com.

Introduction 

The pipeline from drug discovery to development to approval is a complex and lengthy process. However ML is beginning to show innovations in all stages of drug development. Target validation, identification of prognostic biomarkers and analysis of digital pathology data in clinical trials are some of the opportunities in which ML can be implemented. 

There are two main techniques used to apply ML: supervised and unsupervised learning. Unsupervised learning is a type of algorithm that learns patterns from untagged data. Supervised learning on the other hand, is a type of algorithm formed from labeled training data which consists of a set of training examples.

Supervised learning methods have been used to predict future values of data categories or continuous variables. Unsupervised learning is primarily used for exploratory purposes in the development of models to enable data clustering in a format not specified by the user. This particular technique helps to identify hidden patterns within input data, whereas supervised learning methods predict future outputs based on a trained model of known input and output data. According to a 2020 review, supervised learning techniques such as Support Vector Machines, deep learning and regression methods have already been applied to biomedical challenges in the last decade.

Applications in drug development 

Target identification and validation

The identification and validation of a therapeutic target requires the analysis of vast datasets. Genetic screening and high-content imaging are examples of techniques that produce large datasets that can be exploited for early target identification and validation. However analysis of such data requires appropriate mathematical methods to construct valid statistical models – this is where ML can be exploited. 

As early as 2010, ML was applied in a study for target validation in the form of a “decision tree-based meta classifier”. In this study, the ML platform was proposed as a computational approach to predicting morbid and druggable genes. Morbid genes with mutations are associated with causing hereditary human disease. The tree-based meta-classifier was used to predict targets on a genome-wide scale. It managed to correctly recover “65% of known morbid genes with a precision of 66% and correctly recovered 78% of known druggable genes with a precision of 75%”. The ability of ML to reliably predict specific genes on a genome-wide scale is a huge step forward in further optimising target identification. Prediction of therapeutic targets saves time and resources for pharma companies and potentially utilise the mathematical approach to predict more reliable targets. 

Drug discovery

The Generative Adversarial Network (GAN) is an example of a recent innovation in deep learning for drug discovery. Deep learning is a specialised area of ML that attempts to model abstraction from large-scale data using multi-layered deep neural networks (DNNs). Abstraction is a computer science term that refers to the process of filtering out irrelevant data in order to focus on the desired information. 

As an unsupervised ML method, GAN has proven to address the challenges of supervised ML, primarily the training of large data sets which is often expensive and time-consuming. In a 2017 study, GAN-based frameworks were used to develop and identify novel compounds for anticancer therapy with chemical and biological datasets. 

This study emphasised how the productivity of pharmaceutical research is limited by inefficient early lead discovery processes. It also highlighted how in silico-based approaches like deep learning models can generate reliable data at a reduced cost and time scale relative to current screening methods. 

Computational pathology 

In research, a pathologist interprets the presentation of tissue/cells within a glass slide. The spatial context between cells, size and general cellular structure can be indicators of changes with drug interaction. Computational pathology is becoming an important part of drug development. It has been suggested that this method could allow pharmaceutical companies to discover novel biomarkers and generate them in a more precise, reproducible and high-throughput manner. 

ML allows for high-throughput generation of features for thousands of cells, which is an impossible task for pathologists. Immuno-oncology is a particular therapeutic area which has benefitted from using computational pathology. A 2017 study found that computational analysis of tumour-adjacent benign tissue in prostate cancer revealed information typically ignored by pathologists but has been associated with progression-free survival.

Ongoing challenges in adopting AI/ML

One of the main concerns with ML predictions is overfitting or underfitting. Overfitting is described as a model which consists of “lower quality information/technique but generates higher quality performance. In contrast, underfitting models fail to recognize the data sets’ underlying trend and generalize the new data inputted”. Both errors produce inaccurate results which compromise the reliability of predicted drug targets. Increasing the sample size and cross-validation are often used to address these problems. Cross validation is a technique that uses independent data sets to estimate the accuracy of ML algorithms’ models. 

Another challenge for the pharmaceutical industry is the lack of personnel to operate AI/ML-based platforms. Furthermore, there is often skepticism about the quality of data generated by AI. Small organisations are often limited in their budget so cannot afford to invest in AI/ML technology. 

Despite the improvements needed to refine ML applications, the potential they bring to drug development is significant. In addition to reducing human error, the automation of ML software can analyse data from many sources more accurately and in a shorter period of time. The advancement of AI and ML will continue to reduce the challenges faced by the pharmaceutical industry.

To discuss these topics further with sector experts, and to ensure you remain up-to-date on the latest in clinical development, sign up for Proventa International’s Bioinformatics Strategy Meeting, set for 1 July 2021.

Charlotte Di Salvo, Junior Medical Writer
Proventa International

The post Implementation of Machine Learning in Drug Development appeared first on Proventa International.

Companion diagnostic tests are an important part of precision medicine in oncology. In addition to matching optimal treatment options for patients, these tests enable clinicians to monitor the efficacy of therapeutic intervention in disease management. Innovations in cancer biopsies and blood tests aim to introduce more accurate and less-invasive companion diagnostics.

To discuss these innovations and more with other leading experts in an informal setting, sign up to Proventa’s Oncology Strategy Meeting, held online on 17 June 2021. 

Introduction

According to the National Cancer Institute, companion diagnostic (CDx) tests are used to match appropriate treatment for patients with cancer. The test can identify specific genetic changes in tumours or biomarkers that are targeted by a specific drug. In addition, the tests can be used to determine whether the patient is a candidate for treatment or not. Companion diagnostics are also used to monitor the clinical response to specific drugs which helps to ensure safety and efficacy according to the FDA. 

Companion diagnostics are an important tool for precision medicine in oncology as a strategy to optimise efficacy of patient treatment. Cancer pathology is highly complex: with over 100 variations of the disease and differing progression, personalised medicine is greatly sought after. 

In companion diagnostics, mechanisms in molecular genomics (molecular interactions within the genome) are most associated with pathology, and as such are disease targets. Circulating tumour cells (CTCs) and circulating free tumour DNA (ctDNA) are the two main biomarkers detected by liquid biopsy (the analysis of tumour physiology using biomarkers circulating in bodily fluids). ctDNA is becoming an increasingly popular biomarker for a variety of tumours. 

The presence of ctDNA arises when tumour cells release DNA into surrounding tissue, either via apoptosis (programmed cell death) or active secretion. The ctDNA is then delivered into the bloodstream. Circulating ctDNA is then identified by specific genetic mutations for a tumour type. Quantifying the biomarkers present measures the effect of treatment relative to disease progression, hence is an important part of treatment choice and disease management. 

Barriers to use

The delivery of CDx tests is a topic under debate. While invasive diagnostic tests are potentially more accurate as they involve directly testing a patient’s tissue, they are often unpleasant and sometimes painful procedures. Hence, there is a clinical need to develop less invasive alternatives which are widely available and match the accuracy of invasive techniques. 

Unfortunately, there are a multitude of barriers that prevent CDx tests being used in clinical practice. According to a recent article, inefficient market development planning for CDx tests prevents approximately 50% of cancer patients from receiving the correct test for “biomarker-guided treatment”. Other factors include that the test is not always required, and a sample is not always available. These challenges will need to be addressed to allow the right patients to benefit from some of the new tests that have recently been developed. 

Latest innovations in next-generation techniques 

Blood samples over invasive procedures

In March 2020, the IMvigor011 study reached phase III of an oncology clinical trial using Signatera, a companion diagnostic created by Natera. Natera is a clinical genetic testing company that specialises in cell-free DNA testing technology. Their focus is primarily on women’s health, organ health and oncology. In an article published in early 2020, it describes the randomised clinical trial, sponsored by Genetech Initiate, which was launched to evaluate the safety and efficacy of adjuvant treatment with the PD-L1 inhibitor. Eligible patients are to be screened with Signatera, a companion diagnostic used to identify muscle-invasive urothelial carcinoma. 

Signatera is a customised ctDNA test that monitors treatment and also assesses molecular residual disease (MRD) in patients with previous cancer diagnoses. According to the article, the companion diagnostic requires only a blood sample with blood tests that are personalised to each individual relative to the genetic signature of mutations found in a tumour. 

This is an example of the exciting developments in invasive companion diagnostics techniques which maintains accurate detection of disease as well as creating a procedure better tolerated by patients. This particular diagnostic test has been used to “detect recurrence earlier and to help optimize treatment decisions.” The test awaits FDA approval.

Liquid biopsies and next-generation sequencing

In August 2020, the FDA approved the first liquid biopsy companion diagnostic – Guardant360 CDx assay. According to an FDA press release, this test uses next-generation sequencing (NGS) to recognise patients which present specific mutations of the epidermal growth factor gene. This specific type of mutation is expressed in a deadly form of metastatic non-small lung cancer. It is the first approval for NGS and liquid biopsy in one diagnostic test.

NGS is a DNA-sequencing technique which can sequence the entire human genome in a day. According to a review of NGS platforms and applications, “each of the three billion bases in the human genome is sequenced multiple times, providing high depth to deliver accurate data and an insight into unexpected DNA variation”. In the context of companion diagnostics, NGS is used for high-throughput tumour profiling. In one test, NGS allowed clinicians to detect mutations in 55 tumorous genes, instead of assessing one gene at a time. 

In addition to utilising NGS, the Guardant360 CDx assay also uses liquid biopsy. In comparison to solid biopsies, the procedure is less invasive for patients and can also be repeated easily, which could have a considerable number of clinical applications. 

It appears that companion diagnostics are diversifying. In addition to their application in matching therapy to patients, they are becoming increasingly used to identify the presence of residual disease in order to intervene with treatment before it is incurable. This was a point raised in the 2020 Nature article, in which the author states that “during treatment, regular liquid biopsies could reveal the persistence or increase of CTCs or ctDNA — which would indicate resistance to the chosen therapy. People could then be offered a more effective treatment before the tumour burden becomes excessive and incurable”.

To discuss these topics further with sector experts, and to ensure you remain up-to-date on the latest in clinical development, sign up for Proventa International’s Oncology Strategy Meeting, set for 17 June 2021.

Charlotte Di Salvo, Junior Medical Writer
Proventa International

The post Strategies to Optimise Companion Diagnostics in Cancer appeared first on Proventa International.

Given the long-standing hype around AI and machine learning (ML) in pharma R&D, it can be easy to overlook innovations that potentially promise just as much for the pharmaceutical sector. Quantum computing is one area expected to have far-reaching impact within the field, but is at the moment almost singularly misunderstood. 

To find out more about this technology with such great promise, Proventa spoke to experts Dr Emir Roach, Senior Director of Strategic Programs and Head of Emerging Technologies at Takeda, and Dr Philipp Harbach, Head of In Silico Research at Merck KGaA, Darmstadt, Germany, about the future of quantum computing and what lessons the pharmaceutical industry should be learning right now. 

The Distinction Between Quantum and Classical Computing

Quantum mechanical methods have been around for almost a hundred years, with recent advancements in computer hardware making it possible to extend these methods – including combinatorial optimization, differential equations, linear algebra, and factorization to classical computers which can simulate quantum systems in real-life applications. These simulations are already being run in most pharma companies by computational chemists, and applied to day-to-day problems. The problem with this scenario is that running quantum mechanics on classical computers is enormously costly. 

To overcome the issue of costly computations, the use of ‘qubits’ in quantum computers might be a solution. While classical computers use bits with a value of either zero or one, quantum computers use qubits with superpositions allowing a much wider range of possible states than a simple binary. These qubits all act as a group (entanglement) to achieve much higher information density than a regulatory computer, speeding up computation by a vast margin.

It is important to understand that quantum computers will not replace classical computers: they are used for separate issues, not simply improving on current technology – though quantum computers may have an advantage in optimisation and ML problems. They might also have an advantage over classical systems in their ability to tackle quantum simulations, as they are already inherently quantum.

One benefit of quantum computers over alternatives, as Dr Philipp Harbach, Head of In Silico Research at Merck KGaA, Darmstadt, Germany, pointed out, is that when it reaches cost-efficiency the new technology will also be highly environmentally sustainable, as quantum computers need almost no energy compared with a classic computer. 

What is Quantum Computing?

Dr Roach set out roughly how a quantum computer is designed to work: 

“Fundamentally, when you’re performing a chemistry computation on a traditional computer, you’re defining a math problem that the computer needs to solve. The computer pushes bits through arrangements of gates which enable performing calculations and provides the results. 

“With quantum computers, at the high level, you are attempting to represent the initial state of a problem, for example a chemical structure in qubits which can be made to be interrelated – or entangled. Then, you manipulate those qubits using radio waves, light etc. Depending on the arrangement and interrelations of the qubits, the system undergoes changes. Then you read the latest state of the qubits to see where things landed. In effect, you’re creating an analogous representation of the chemical and manipulate it to observe changes.

“What that means is, for every problem, we need to figure out what it means to create that initial state of qubits, figure out how to build bridges between those qubits, what it means to fire light/radio waves into those qubits, and then figure out after we’ve done a reading of the new state of those qubits, how to interpret them. It’s a different translational discipline.”

Quantum computing is a natural fit for the pharmaceutical sector. The field’s need to model molecules and determine their properties can be greatly enhanced by quantum computing, which will hugely increase the ability to understand the effect of a given molecule on the body through simulations of its mechanics and properties. Use cases of quantum computing in pharma can extend drug development however, into supply chain, clinical trial scale-up etc.

Current Challenges Facing Quantum Computing

This ‘quantum advantage’ is currently muzzled by a number of complex challenges. First and foremost, as Dr Harbach noted, there is no existing solution yet for scaling the hardware of quantum computers while also reducing the exponential problem of noise (e.g. error) that increases too. 

The next challenge is simply that currently, few companies know what to do with quantum computers. While theoretical algorithms exist for the technology, they’re few in number and require complex and expensive hardware. In contrast to this, quantum methods on classical computers date back several decades: there are thousands of algorithms that can be applied. 

Dr Roach set this out more clearly: “Once we have a problem in mind, translating that into a quantum algorithm to run on a computer is challenging. That translation is something for which there is not much expertise in the world.”

Dr Roach expounded on this problem by pointing out that for most current problems within pharma R&D, a solution already exists besides quantum computing: “There are many models out there that solve current problems, from finding the right target, to generating a molecule and observing its interactions with the targets.”

“We as an industry need to zero in on the most pressing, highest-value problems we want a quantum computer to solve. There are some high-level understandings of where we want to go, be it protein-folding or generating drug designs, but no clear view as of yet.”

These problems are not, however, simple binary issues that will usher in a new era of quantum computing when overcome. Each type of challenge – hardware, expertise etc – contains categories of difficulties within them. Within the hardware problem, there are architectural challenges (for example, how qubits engage with experts’ designs) which influence the type of computations which can run on the hardware. 

Within translation, smaller challenges exist around how to initialise a quantum computer, and then how scientists should read outputs and feed them back into the quantum model. 

The Quantum Advantage

According to Dr Harbach, a rough estimation of the time quantum computing would gain an advantage over classical technology somewhere in the pharma product life cycle – not necessarily just in R&D – would be around five years from now. This still would not be cost-efficient, however, and would only be used to solve extremely complex, vital problems. To reach a mass market level, Dr Harbach estimated that perhaps ten to 15 years would need to pass, though this is by no means certain. 

Lessons for the Pharma Industry

The good news is that the industry is working as one to solve the quantum issue. While big headlines in recent months – the Boehringer / Google development, Roche’s partnership with UK firm Cambridge Quantum Computing – have made the area seem an increasingly heated competition, in reality pharma companies are working together in this pre-competitive phase, and do not expect real value on short-term, according to Dr Harbach. 

Most companies in the pharma sector are connected by QuPharm, an interest group dedicated to sharing research around quantum computing. Dr Roach spoke more about the organisation: “As an industry, I’d argue we recognise that there are many steps left to take before we overcome the challenges of developing quantum computers.

“To get past these challenges will require a number of disciplines to come to together, from hardware specialists and translators who can turn chemistry problems into quantum algorithms to computational chemistry experts who can find where exactly limitations are, through to business strategists who can determine how much should be invested on a problem.”

This cautious, collaborative approach to quantum computing leads back to an earlier point about one of the main challenges in the area: the lack of questions that can be currently answered with the technology, and the lack of a business problem for quantum computing to solve that would reinforce the need to invest in that area. 

As Dr Harbach said, it is dangerous to promise that all current simulation problems or even some business problems can be answered by quantum computers: “This is dangerous, it’s a scorched earth policy. When computer models first appeared in the 1980s, the promise that they would revolutionise pharma development became overhype, and that false promise caused a lot of damage. It is very important to be careful in this regard.”

Joshua Neil, Editor
Proventa International

The post The Promise of Quantum AI in Pharma R&D appeared first on Proventa International.