How is Artificial Intelligence (AI) accelerating the Drug Discovery Process?

Artificial Intelligence (AI), as the name suggests, is intelligence demonstrated by machines. It has human-like cognitive behavior and thus has gradually taken over the pharmaceutical sector, in irreplaceable ways. Why I chose to make that statement is because it has been overpowering conventional means of drug development (the cynosure of the pharmaceutical industry) at a temporal scale over the past decade. Albeit not at the cost of human intelligence, AI employs tools that improvise human inputs intelligently to generate an output that holistically makes the process of cognate drug discovery less cumbersome and more accurate.

Drug Discovery Process

The ultimate goal of drug discovery is to identify chemical entities that can be used as therapeutic agents to combat diseases. However, this entails a list of chronological steps until the drug finally qualifies from the bench to bedside! Briefly speaking, they can broadly be enlisted as:

  1. Basic Research, begins in the laboratory (traditionally speaking)
  2. Preclinical Research, safety check on animal models
  3. Clinical Research, check on human subjects
  4. Food and Drug Administration (FDA) review, the final Yes or No
  5. Post marketing review

Where does Artificial Intelligence come in Drug Discovery ?

AI accelerates Step 1 right away! It expedites the process of drug design using various tools before, during and after hands-on research. In April of this year, a German biotech company in association with Exscientia, Oxford announced phase 1 clinical trial of a novel anticancer molecule. Exscientia, which uses AI, accelerated the process several folds from 4-5 years to just months saving not just time, but billions of dollars. Its application can be broadly classified as:

Need Counseling?

Book a 1-on-1 session with Diya

Target Selection

AI analyses drug databases from public libraries for predicting their therapeutic potential. By using deep autoencoder, relief algorithm and binary classification it enables target prioritization. This involves structure prediction and its probable effect in biological milieu.

Screening and Lead Optimization

Screening for a drug and final optimization requires several Hits for final Lead Identification. AI-based virtual screening which skims through millions of related compounds and their structure utilizes automated chemical synthesis, image-activated cell sorting devices to speed up the individual tasks.

Preclinical Studies

Pre-clinical studies are performed in the laboratory to establish its efficacy and safetyin vivo. Machine Learning (ML), a subset of AI exploits pharmacokinetic/pharmacodynamic modelling algorithms to predict toxicology profiles via optimal dose concentrations in drug-dose response. Deep Learning (DL) scripts are used in ‘in-silico’ methods under various biological systems and conditions to gauge pharmacological properties by implementing decision-making algorithms.

Clinical Trials

The use of AI in this step is a little dicy! The biggest challenge here is to answer questions on computational prediction because human subjects are involved.

Top AI-based Drug Discovery Startups/Companies

The customary trial-and-error method of novel drug identification costs companies both time and money, two of the most important parameters anyone would not want to compromise with. With the advent of AI and its subsets like ML, DL, the whole process has been scaled down as far as both time and money are concerned. Hence pharmaceutical companies have collaborated with several enterprises for better outcomes (qualitative and quantitative) over the past decade. The top AI-based drug discovery start-ups and companies are:

  • Generate Biomedicines (USA) : Protein Therapeutics
  • Turbine Simulated Cell Technologies (Hungary-UK) : Anticancer Drugs
  • Genome Biologics (Germany) : Cardiovascular and Cardiometabolic Therapeutics
  • XtalPi (in collaboration with Pfizer) (USA) : Molecular Therapeutics
  • Insitro (USA) : Biomarkers and Drug Efficacy
  • Cambridge Cancer Genomics (UK) : Precision medicine for Cancer
  • Exscientia (UK) : Candidate Drug Screening

Career Transition(s) from Pharma/Biotech to Data Science/AI

While a background of Mathematics with Computers is often considered apt for a career in AI, biotech graduates can also switch to data science and AI with basal knowledge of coding. This, nonetheless, is dependent on the subject combination with a positive bias towards having Mathematics and Computers at the undergraduate level. To know more about career transition from biotech to AI do give this a read to know more.

Future of AI in Drug Discovery

AI in Drug Discovery is time and cost-effective, together with a higher level of accuracy (computational methods employed) and hence success-driven. Because drug discovery (and development) requires various cross-productivities, the application of AI is all the more pronounced. Although there are several challenges that need to be worked on (and are being troubleshooted), the future, nonetheless, is bright! The prospects of AI in drug discovery is plummeting as it is dominating in the post-discovery chain of events (implementation per se for marketing) as well, namely:

  • Drug Repurposing
  • Polypharmacology
  • Off-target Effects
  • Quality Control and Assurance

Related Articles :

Computational Biology Careers and Jobs in India

Current Artificial Intelligence Trends and its Future post-COVID-19 (Jobs & Salaries)

Featured Image:

Translate »
%d bloggers like this: