Immunotherapy Development - A bench-bedside case study

Author: Dr Craig Davison, Sonrai Discovery Partnerships Lead at Sonrai Analytics

Updated: 05/04/2024

Dr Craig Davison

Dr Craig Davison

Sonrai Discovery Partnerships Lead at Sonrai Analytics

Craig joined Sonrai Analytics as Sonrai Discovery Partnership Lead in July 2023, bringing with him a wealth of expertise in preclinical R&D and translational medicine. In his current role, he aids prospecting clients to decide if Sonrai Discovery is the right solution for their challenges through engaging discussions and bespoke live demonstrations. He is also responsible for producing accurate and engaging marketing material to showcase Sonrai’s wealth of capabilities. Additionally, he works closely with the product team to understand the product roadmap and to shape this based on feedback from prospective clients and developments in the world of precision medicine. Before joining Sonrai, Craig worked as a Scientist jointly within academia/industry at Queen’s University Belfast and CV6 Therapeutics. He earned his PhD in Molecular Biology and has amassed 9 years of experience in precision medicine research.

Data types in this case study: NGS Screen, In Vitro Assays, In Vivo Models, single cell RNAseq, Bulk RNAseq, Proteomics, Clinical trial


This use-case showcases some of the real work being done on the Sonrai Discovery Platform by our clients. It utilizes publicly available data from a variety of sources (see table 1) that has been manipulated to create a unified story of preclinical drug discovery and development all the way to clinical deployment.

The primary objectives were to showcase the ability of the Sonrai Discovery Platform to manage and analyze data across the drug discovery and development process. Secondary objectives were to showcase how effective data analysis, visualization and the use of cutting edge machine learning (ML) techniques can significantly enhance drug discovery and the development of therapeutics with high likelihood of success.

The Difference Sonrai Makes

Every organization is unique, and generic solutions fall short in addressing their distinct challenges. Sonrai is purpose-built to tackle the complexities of precision medicine, offering tailor-made solutions that align with your specific needs.

Discover Predictive Biomarkers 52% Faster

Sonrai’s AI technology supports rapid biomarker identification, empowering our partner to predict optimal therapy responses through an innovative toolkit.

57% Acceleration in MoA Studies

See how Sonrai unified diverse data sources to rapidly understand complex molecular pathways, helping identify predictive biomarkers and drug mechanisms of action.

14% Reduction in Preclinical Phase

Discover how Sonrai helps pharmaceutical companies overcome challenges in Invivo Drug Efficacy with AI-powered analytics.

Jump To IO Development Stage

Figure 1: Project overview page which links all the data, team members and analysis in one place

Novel Immuno-Oncology Antibody Discovery

Our client required help to develop novel antibodies that showed strong target binding affinity and specificity. An NGS antibody screen was used, alongside ML in silico methods, to identify antibodies with strong binding potential for the target. The Nextflow airrflow (1) pipeline was used to process raw FASTQ files within the Sonrai Discovery platform (Fig. 2-3). Following screening, antibodies that were selected for their binding potential (Fig 4) were assessed for their characteristics. ELISA affinity assays were utilized to identify antibodies with the strongest binding that could be taken forward for further preclinical development (Fig 5).

0 Novel Antibodies Identified

with Strong Binding Potential

The number of novel antibodies identified with strong binding potential was 50, showing an increased output and potential for higher success rates in finding viable therapeutic candidates.

0% Efficiency Rate

of the NGS Antibody Screen in Identifying Potential Candidates

The efficiency of the NGS antibody screen in identifying potential candidates achieved an 80% efficiency rate, showing a high success in validating candidates through subsequent assays, indicating a robust screening process.

0% Time Reduction

Antibody Discovery Process

Time saved in the antibody discovery process through ML-based in silico methods was a 50% reduction in time, demonstrating the impact of machine learning in speeding up the discovery phase.

Do These Challenges Sound Familiar?

Contact our friendly team for expert guidance and transformative insights.

In Vivo Preclinical Development

In the second stage of their novel therapies development, we utilized in vitro and in vivo experiments to determine that their novel immunotherapy worked best in a particular cancer type and in combination with two other standard-of-care treatments. Figure 6 demonstrates that the novel triple combination induced complete responses in vivo. Multiple assays were used to determine potential mechanisms of action for this novel combination, including digital pathology of tumor samples and single-cell RNA-Seq data. Using no-code data visualizations and analysis within the platform we were able to determine immune cell population changes induced by the treatment (Fig. 7). Of note, the triple combination treatment induced increased activated immune cells over time, for example activated NK cells (Fig. 8) and CD4+ T cells. Our data analysis provided clear evidence for potential treatment efficacy and elucidated mechanisms of action.


Complete in vivo response rate

Percentage of tumor volume reduction in response to novel immunotherapy: A novel immunotherapy given in combination with standard-of-care treatments led to a 100% complete response rate in vivo.


Faster time to insights

In reduced project timelines, decreased labor needs, and fewer repeat experiments. These savings and faster time-to-market for drug development substantially increased the return on investment.

Clear Evidence of Therapies MoA

Data visualization and analysis through no-code applications allowed rapid interrogation of immune cell population changes in response to treatment, resulting in clear evidence of the therapy’s mechanism of action.

Figure 6: Line chart analysis revealed that their novel treatment combination outperformed all other evaluated treatments

Figure 7: Single cell RNA-Seq analysis of circulating blood cells revealed treatment induced immune cell changes

Figure 8: Box plots demonstrated treatment induced immune cell changes over time

Investigate Target Expression

When clients are developing therapeutics against novel targets, there is frequently a challenge to determine possible side effects. Sonrai developed the below ‘Weaver Plot’ tool. It lets researchers integrate and analyze gene data from multiple studies in a single view, allowing them to clearly understand expression across different tissues and disease indications. This has proven to be a great tool to identify successful targets with limited off-tumor effects better. This can be done with publicly available data, as shown in the video, and proprietary data, for example, for a particular disease indication. This tool has led to breakthroughs, advancing targets to clinical trials with a clearer safety profile. Thorough assessment of on-target and off-target expression provided evidence of therapies on target effects that were critical for a successful IND submission.

0 Targets

with enhanced safety profiles

Identification of 5 potential targets with enhanced safety profiles. Successful targets were identified with limited off-tumour effects.


Target expression

Understanding target expression across different tissues and disease indications has improved twofold. This is crucial for developing targeted therapies with fewer side effects.

0% Reduction

Target evaluation

The time and resources spent evaluating targets using the Weaver Plot tool were reduced by 60%, significantly optimizing the target validation process.

Early Clinical Trial

Another client had multiple assets in various clinical stages. They needed to compare their novel therapies with standard of care, identify responders and non-responders using Response Evaluation Criteria in Solid Tumors (RECIST) criteria, and monitor therapy safety. This data was brought into the platform from multiple clinical trial sources in real-time and was analyzed using no-code applications to instantly visualize the longitudinal data. Figures 9-10 showcase longitudinal survival analysis and figure 11 showcases a dashboard for monitoring adverse events. Previously this client had to request and pay per report to evaluate ongoing trials, which could take days to arrive. Now they are able to evaluate data as soon as it is updated in the platform and they have the ability to version control their datasets, ensuring data integrity is always maintained.

0 %

Higher response rates

Comparative analysis of novel therapies with standard of care using RECIST criteria demonstrated 38% higher response rates, indicating more effective treatment options.


Survival rate improvement

Survival rate improvement demonstrated through longitudinal survival analysis saw a 23% improvement, directly impacting patient outcomes positively.

Days to Real-Time

The time spent evaluating ongoing trials was reduced from days to real time. This enabled immediate access to trial data, which sped up decision-making processes.

Figure 9: A survival plot demonstrating a novel treatment’s superiority over standard of care treatment

Figure 10: A survival plot demonstrating differential survival in patients categorized based on RECIST criteria

Figure 11: Plots monitoring adverse events in treated patients to evaluate safety profiles

Prognostic Biomarker Development

During the clinical development of a novel treatment, it became clear to one client that they needed prognostic biomarkers to stratify patients to improve their odds of success in their later-stage clinical trials. With that in mind, tumor and blood samples collected before treatment generated various omics datasets. Below is an example of a bulk RNA-Seq dataset used to identify potential biomarkers of response and resistance. Both traditional differential expression and ML methods were used to identify biomarkers of interest. Using no-code applications to accelerate this discovery effort, their R&D team were able to discover clinically relevant biomarkers within a single day of the omics dataset reaching the platform. This collaborative research environment accelerated this biomarker discovery process by months.

An example ML-based workflow below showcases how PCA analysis was used to visualize this complex dataset containing thousands of potential biomarkers and identify that there are distinguishing features between responders and non-responders (Fig. 12). An XGBoost classifier model was then trained to distinguish responders and non-responders (Fig. 13-14). This model was used to identify important features that could be individually assessed (Fig. 15-16). These were then taken forward for further clinical development to create a prognostic algorithm capable of stratifying patients for their novel therapy.


Clinically relevant biomarkers identified

The clinically relevant biomarkers were identified within a week, accelerating the pace of biomarker discovery.

0% Accuracy

XGBoost classifier model classifier

In distinguishing responders and non-responders, enhancing and accelerating the precision of patient stratification.


Biomarker discovery time reduction

Leveraging Sonrai ML-based workflows reduced the time spent on biomarker discovery processes by 72%, showcasing the efficiency of ML in streamlining research processes.

Figure 12: PCA analysis was used to reduce the dimensionality and visualize this complex data

Figure 13: An XGBoost classifier model was trained to distinguish between ‘Responders’ and ‘Non-Responders’

Figure 14: Classifier model performance can be tuned and evaluated for rapid development and discovery

Figure 15: XGBoost classifier within the platform provides features of importance allowing for ML-based discovery

Figure 16: Box plots were used to investigate individual biomarkers for differential expression


The Sonrai Discovery platform has been used for all stages of drug discovery. This case study covers some of the data types and analytical methods frequently performed on the platform. With the Sonrai Discovery platform specifically designed  for precision medicine, it can handle all the required data types needed for drug discovery and development. Purpose-built platforms, like Sonrai Discovery, have the potential to revolutionize drug discovery and development, improve the success rates of clinical trials, and ultimately impact patients through better use of data.

Access to data and the ability to analyze that data with cutting edge ML tools, should no longer be a blocker to clinical development. If you want to see the Sonrai Discovery platform in a live demo, reach out to our team and we can organize a meeting for you. With in-house expertise across a wide range of data types and analytical techniques, our team will be happy to discuss your projects and requirements and see if the Sonrai Discovery platform is right for you and your team.


  1. The nf-core framework for community-curated bioinformatics pipelines. Philip Ewels, Alexander Peltzer, Sven Fillinger, Harshil Patel, Johannes Alneberg, Andreas Wilm, Maxime Ulysse Garcia, Paolo Di Tommaso & Sven Nahnsen. Nat Biotechnol. 2020 Feb 13. doi: 10.1038/s41587-020-0439-x.

Table 1: Data used and source

Use Case




Novel IO antibody discovery

NGS Antibody Screen

  • NGS data processing
  • In silico binding prediction
  • Amino acid frequency calculation

Synthetic data representative of clients use-case


  • Binding affinity
  • EC50 calculation

In vivo preclinical development

In vivo tumor volume experiment

  • Tumor Volume analysis
  • RECIST analysis to compare treatments

Synthetic data representative of clients use-case

In vivo Single Cell RNA seq

  • Immune cell counts
  • Immune cell proportion comparison by treatment and time

Investigate target expression

Multiple RNAseq and proteomics datasets

  • Algorithm that…
  • Comparison of target expression by tissue

See ‘weaver plot’ video

Clinical trial analysis

IO clinical trial outcome data

  • Survival and outcome comparison 

Modified from;

Riaz –

Allen –

Hugo –

Nathansan –

IO clinical trial RNA seq data

  • RNA seq data processing
  • Volcano Plot
  • XGBoost
  • Pathway Analysis

Discover how companies worldwide grow with Sonrai. Explore all our case studies.

Advanced Data Management and Analytics in Clinical Trials

Explore advanced data management and analytics in clinical trials with a synthetic Phase I trial of Tyzolimab for Advanced Esophageal Squamous Cell Carcinoma (ESCC).

Enhancing Invivo Drug Efficacy

Discover how we help pharmaceutical companies overcome challenges in Invivo Drug Efficacy with AI-powered analytics.

Integrating Imaging and Omic Data

Explore how Sonrai's advanced analytics can help companies overcome challenges in drug development by integrating imaging and omic data.

Get in touch

Like What You See? Let's Talk

We Listen to your problems
We give you confidence to make your decision