Infinite Curationインフィニットキュレーション最先端コンピュータ技術の最大活用により「ミクロ」遺伝子解析と「マクロ」医療画像解析の相乗から新たな価値を創出し、医療・ヘルスケアの進展に貢献します。

Contracted or co-development of advanced functionality [development cases]

We are a medical research and development company.
We are expanding our accumulated development experience in contract research for research institutions and next-generation CT contract development for major medical device manufacturers.
We have a team of experienced medical engineers, and we can flexibly respond to your development requests in an all-round way by using processors developed by our group companies and by utilizing high-speed technology.

 

  • Advanced
    Technology
    Development
  • One-stop,
    wide-ranging
    support
  • We listen to
    your requests
    and respond
    flexibly.
  • Short-term
    development
    by highly skilled
    engineers

 

For example, we can respond to the following developments
  • ・Incorporate specific features,such as AI or 3D processing
  • ・Automation, faster processing
  • ・UI design of the front part
  • ・Make your program web-enabled or cloud-enabled
  • ・Back-end server development
  • ・Integrated development of the entire application

Strengths of Infinite Curation’s Development

1.One-stop development of medical equipment, medical imaging systems, and medical information servers

We have engineers with extensive experience in a wide range of image processing and medical information processing, and our expert staff can develop products in a short period of time.

We provide one-stop contracted development, joint research and development, consulting services, etc., for everything from image reconstruction engines, which are the main components of medical equipment devices, to operation interfaces, 2D/3D/4D displays and various types of analysis in post-processing, and external use on image storage and distribution servers.

For example, we will respond to a wide range of development requests, such as medical AI workstations for multi-purpose use and medical servers for cloud computing.

2.Parallelization and acceleration for high processor performance

We have experience with high-speed arithmetic processors and NVIDIA GPUs, as well as advanced software technologies and processing acceleration technologies, including software implementations for deep learning.

We will leverage this strength to parallelize and accelerate large-volume medical image processing.

3.High-speed processing technology in the field of genome analysis

Based on the metagenome analysis using shotgun data on supercomputers at research institutes and other institutions and on the demonstration of human genome analysis technology, we will work to achieve high-speed analysis of large genome data, which is expected to grow in size in the future.

4.Expand utilization of supercomputers and HPC technology

We will reflect supercomputer and HPC technologies in the medical and life science fields and expand our business by leveraging our accumulated experience in supercomputer utilization.

DEVELOPMENT CASE:
Human Whole Genome Analysis System
Developed system example: ZettaScaler-3.0 Server Unit for Human Genome Analysis
To achieve the world’s highest accuracy and speed, we are continuing to develop a whole human genome analysis system in cooperation with PEZY Computing Inc. and ExaScaler Inc.

1. Overview and Features:

ZettaScaler-3.0 Server Unit provides ultra-fast secondary genomic analysis of whole human genome sequencing data with high accuracy. The server equips with four PEZY-SC3 processors, which were originally designed and developed by PEZY Computing group.
・A single server unit can process about 100 human whole genome data (equivalent to 100 Gbp/sample) per day.
・It provides results with extremely high analysis accuracy without sacrificing accuracy for speed.
Example: SNP F value: approx. 0.999, INDEL F value: approx. 0.996, processing time: approx. 15 min (100Gbp conversion) (Genome In A Bottle, HG001 benchmark ver. 3.3.2 was used for the evaluation.)

 


 

2. Analysis Workflow:From FASTQ file input to VCF file generation

[FASTQ] → Alignment → Coordinate Sorting → Mark Duplicates → Base Quality Score Recalibration & Apply BQSR → HaplotypeCaller→ [VCF]

 

 

Accelerated software used in the GATK Best Practice pipeline, the most commonly used in human genome analysis, completes to variant calls in less than 15 minutes when 100 Gbase FASTQ is used as input.
The following improvements have been made to increase speed
1) Acceleration by PEZY-SC3 porting
2) Intermediate files made on-memory
3) Optimization of CPU processing

We are not only increasing speed, but also improving accuracy.
1) BWA MEM has been implemented with options to improve accuracy.
2) Highly accurate probabilistic model implemented in GATK 4.2 is implemented.

 

3. ZettaScaler-3.0 Server Unit specification

baseboard: ExaScaler EPX-BASE2 x 1 CPU: AMD EPYC x 1
accelerator:PEZY Computing MOD-SC3H (PCIe Gen4x16 bus) x max 4 modules
main memory:DDR4 ECC Registered 3200MHz SDRAM max 1TB (64GB DIMM x 16)
storage:M.2 NVMe SSD 2TB x 4 OS:CentOS

 

4. PzBWA-MEM

PzBWA-MEM is a fast alignment software based on BWA-MEM version 0.7.17 (r1198) 1 with improvements by PEZY Computing.
1) Acceleration of alignment process by PEZY-SC3
2) Optimization of the number of pipeline stages and pipeline structure for faster processing
3) Fast query data loading by optimizing Fastq loading
4) Faster post-processing by making the output on-memory
5) Add options to adjust scores, etc. (By default, it works the same as BWA-MEM.)
6) Lift-over function of alignment to Alternate contig.
7) Sensitivity improvement using information on known mutations

 

5. PzHaplotypeCaller

PzHaplotypeCaller is a software for genome mutation analysis based on HaplotypeCaller of GATK4.2.0.0, which has been made faster and more accurate by PEZY Computing.
The improvements are as follows
1) Full scratch in C++ based on GATK4.2.0.0 HaplotypeCaller
2) Accelerated processing using PEZY-SC3
3) Optimization of CPU processing for higher speed
4) Probabilistic model implemented in GATK 4.2
・Foreign Read Detection (FRD)
・Base Quality Dropout (BQD)
・DragSTR

 

DEVELOPMENT CASE:
Medical image 3D analysis viewer with GPU-based real-time rendering

It is implemented in the CUDA programming environment and can run on a wide range of PCs, from notebooks to large servers, using NVIDIA GPUs.
The program has been improved to read and display in 3D a 1000-slice data file of CT images in a few seconds. Volume rendering with real-time processing and still high image quality is realized.
Below are several example images from the 3D Analysis Viewer.

 

The following images show an example of screen layout and vascular analysis. The top left image is the basic 2×2 layout. The other images are vascular analysis images. The red arrow in the lower right image indicates the location of coronary artery stenosis. You can choose from multiple screen layouts.

 

 

The following image is an example of a virtual endoscope display. On the left are multiple images of the colon using CT data. On the right are multiple images of the bronchus using CT data. In addition to manual running, the system can detect the centerline of the colon or bronchus and automatically run along it in real time.
Image A, below right, is an example of transmitting the bronchial wall and observing the condition of the surrounding tissue.

 

 

The following image is an example of a display image of a head blood vessel. The upper side is an example of multiple displays of head aneurysms using CT data. The distance of the neck of the aneurysm and the height from it are measured. The lower side is an example of multiple display of head vessels using MR data.
The icon list on the left is an example of a 3D color template. Many types are available for each human body part, and the display can be updated to that type with a single click of the icon.

 

 

The following images are examples of 3D views of the dentition, mandible and maxilla by CT. The upper right image is from medical CT data, and the other images are from dental CT (CBCT) data.

 

 

The following image is an example of a virtual endoscopic view inside the mandibular canal, which contains important nerves and blood vessels. In image A, to confirm the location of the tooth root within the solid circle, the upper bone of the mandibular canal is intentionally made transparent to expose the tooth root (within the dotted circle).

 

 

The next image is an example of a virtual endoscopic view in the mandibular canal, similar to the image above. This is data after treatment with implants. In images A and B, the upper bone of the mandibular canal is intentionally made transparent to expose the implant in order to confirm the implant location.

 

 

The following multiple images are examples of straight views and measurements of mandibular canal cross sections. The positional relationship between the mandibular canal and the tooth root or implant can be clearly displayed. The upper right image measures the distance between the mandibular canal and the root of the tooth.