Future Healthcare: Focused on Scope of Clinical Informatics

BY: Rajavarman Kittu - 21 Jul-2021

Modern medicine is a dynamic science which connects the dots between the biomedical and Data science. Apparently, the analytical Informatics skill set and high-end computation is currently in demand not only for analytical Data scientist but also for clinical labs as they handle humungous data, ranging from Tera to Petabytes of clinical data.

For example, the world’s leading genomics player, Illumina’s Novoseq6000 sequencer which can do large scale whole-exome sequencing or larger comprehensive targeted panel can give throughput like ~6000GB in one single batch of run. Over traditional Molecular Biology diagnostics methods, the Next Generation Sequencing [NGS] method is being applied widely in clinical space. The American College of Medical Genetics and Genomics [ACMG] has defined good practice guidelines with standard workflow recommendation and brought the Genomics diagnostics approach to the light for one and all.


However, there are a few bottlenecks which must be taken into account in the context of computation needs of implementation and performance.  

Difficulties observed in Clinical Genomics Healthcare segments are as follows:

1) Implementation and Management: Most of the International level Indian healthcare providers and referencing clinical labs are equipped to store and manage large data sets. High Performance Computing [HPC] facility or data centers or cloud computing might have capacity of handling storage. Building all possible analysis pipelines with respective software, fine tuning optimized parameters, integrating all the relevant databases for annotations are the key challenges. Maintaining version tracks of both software’s and databases relevant to reporting and meeting all the International standard best practice guidelines is not feasible with many of HPC server providers or cloud computing service providers or hardcore Clinical informatics solution providers. Reaching a “gold standard” continues to be a big challenge, because of several complications like, the need for wide range of skill sets, with close monitoring and continuous progress is needed as the field is very dynamic and evolving very faster than we imagine in Medical field.

2) Life or Death situations - Clock is ticking: We have reached a level where multiplexing can be done in a sequencer, high throughput data can be achieved from the largest machine in the genomics market. The instance question from   referring doctors or patients while giving any sample type [Blood or FFPE or Tissues or CVS or amniotic Fluid or any other types] is “How soon can we get the report, as the next plan of action is totally dependent on the based reports?” 
This can be for any patient, say to assess the possible therapy options for a cancer patient, or for someone who is under medication for monitoring purposes. Clinical genomics being well-characterized solutions for many complicated cancers, if diagnosed at an early stage, rather facing life or death situations at final stage of cancer.

3) Performance of Computational Genomics: Refined, optimized algorithms should be in place for achieving short Turn Around Time [TAT] with high standard of good quality”. This kind of reporting with reliability is possible by adopting Field Programmable Gate Array [FPGA] technology. The FPGAs is generally defined as semiconductor devices, based on the matrix of configurable logic blocks, which is programmable. The tasks, workflow and the applications which have to be processed through prototypes is pre-defined and well characterized then FPGA Method is a Master key. Handling tons of samples in batches is the current need in diagnostics industry with the consideration of professional guidelines is feasible with FPGA methods.

“Complicated medical questions can be answered by collective actions with good analytical interpretations and strong computational facility.”


4) Advantage of FPGA over traditional CPU/HPC/Clusters:

Performance: While comparing with other modern Micro-processors, the FPGA’s parallel nature gives high speed processing power with respect to time. This demanding system is expected in Clinical diagnostics field. The Microprocessors in general have limitations in processing speed with respect to the dataflow, these issues are not seen in the current FPGA as it has very much improved performance in place.

Reliable and flexible:  Generally, FPGAs are trust worthy, reliable systems. The FPGA semiconductor chips are very flexible in a way that is easy to both reprogram and reuse. The logic cells in FPGA and their interconnections are so flexible to reconfigure the chips based on the need of optimization.

Maintenance: The upgradation can be done through reconfiguring the program without any high maintenance. When the algorithm and mechanism are placed the investment requirement will only be on the circuit boards and hardware. 

Cost effective: As mentioned before, the reprogram options bring down the cost significantly, as it won’t need any redesign if there has to be any smaller changes to be made in FPGA, therefore, they are very cost effective as well.

“FPGA chip promise to do whole exome sequencing data analysis in 8 mins”


5) Demand for FPGA chips in diagnostics: The multispeciality hospitals with all ultra-modern, cutting-edge facilities in which different types of patients walk in to different specialists for general consultation, possible precaution options, available screening options, actionable treatment options. Having said that, except the “IVD” NGS kits assays used all others kits are “Research Use Only” [RUO]. These RUOs in market have to be validated internally in the laboratory with Medical test accreditations board guidelines to bring to the laboratory standard. 

Whole genome sequencing for germline or somatic screening:  The cancer panels are able to capture all possible hereditary germline or acquired Somatic variants [Single Nucleotide Polymorphisms, Indels, Structural Variants and Copy Number Variants]. While the given clinical samples are with expected number of read depth and full coverage of all regions; the FPGA based methods and its algorithms are fine-tuned to perform efficiently with sensitive quality measures. 

Covid-19 Sequencing for Screening and Surveillance: The next generation sequencing data of patients with COVID-19 virus infections symptoms can be analyzed to know the status of “Positive or Negative”. Also, as a part of the surveillance process the patient samples can be analyzed, checking for all possible variants observed. Moreover, based on the variants the clade classification can be achieved. For India, the number of cases are increasing in each wave (I and II), also it is being noticed that the virus is very much virulent; where the infection rate and death rate are spiking in triple mutant compared to double mutant observed before. As far as the Clinical trials and research is concerned, with respect to bioinformatics, revised functional studies with respect to the latest alterations; revision and refinement of vaccination is needed to ensure the effectiveness on patients with latest infections. Because of the above reasons, it calls for more and more analysis of patient data sets and rechecking in detail to get more insights in order to come-up with better outcomes.    

“Food for thought: It’s been 2 consecutive years of the pandemic; inspite of being one of the most important economies in the global market; Shockingly, we haven’t done enough of covid genome sequencing, yet! When will we have all the Covid sequencing data needed?!!”

Comprehensive Genomics Profiling [CGP]: The FPGA chip can analyze the largest targeted panels to answer the variants [Single Nucleotide Polymorphisms, Indels, Structural Variants and Copy Number Variants] from DNA samples and possible fusions from RNA profile, to get more insights about Oncology biomarkers like Tumor Mutational Burden (TMB) and microsatellite instability (MSI) etc, as it is the need of the hour in oncology segments. 

6) Demand of FPGA chips in Clinical Research:

RNA Sequencing: The expression of genes, quantification of genes, all possible fusions can be analyzed in bare minimum time for a larger set of samples. These pipelines are currently well streamlined so directly heading to downstream analysis like Gene Ontology, Pathway Analysis, Cellular function, Biological components and so on in a shorter time is achievable.

Single Cell RNA sequencing: The Cell imaging (spatial) with gene expression profile is a huge challenge if we need to develop in-house pipeline of your own. Also, to process the large datasets, detailed understanding of downstream processing knowledge in both Molecular biology and Immunology is required along with strong analytical skills. Therefore, the good resolutions, optimized computational facility like FPGA is the ideal choice for large set of Single cell data analysis.

Methylation: For large population sample data sets finding methylation patterns based on the CpG islands to answer the epidemiology aspects with respect to the disease conditions can be achieved through FPGA methods.

“In highly populated countries, to conduct national level population based epidemic studies twin “Dragen Bio-IT Platform” might be needed to take sensitive decisions based on the findings and insights in committed shortest time frames ever.”

7) Important factors of Clinical Genome informatics:

Building data analysis pipeline which consists of 

- Downloading software
- Installing software
- Using samples data sets for validating right software’s
- fine tuning the parameters of different software
- testing them with known and blinded samples
- Streamline with all kinds of different workflows

All the above can be done in inbuild-FPGA with improved speed of process as a add on since is it programmable in chips.

The software and database upgrades for all kinds of NGS applications is much more feasible in FPGA as compared to any other methods like HPC/CPU/Servers.

FPGA provides minimum burden of data processing, while handling large, sensitive clinical datasets to answer the significant findings with all clinical standards and best practice guidelines.

“Smart FPGA methods are the future of Clinical and Research Genomics world.” 

About Author


Rajavarman Kittu
Clinical Bioinformatics Specialist
Premas Lifesciences Pvt Ltd

suggested Tags

Partners & Associates