General

Sentieon has released the DNAscope MGI WES model, assist…

At the 41st J.P. Morgan Healthcare Conference on January 9th, 2023, Complete Genomics, announced the launch of its full line of sequencing products in the United States. At the conference, Dr. Rade Drmanac, Chief Scientist of Complete Genomics, gave a speech introducing a series of products based on the DNBSEQ platform. BGI Genomics will provide flexible and diverse sequencing product options covering low to high throughput to enhance the sequencing capabilities of users in various fields and promote the translation of scientific research and applications. In addition to the four sequencing instruments that have already been shipped or are about to be launched, Dr. Drmanac also emphasized the importance of gradually establishing an open and win-win industry ecosystem and long-term cooperation between industry, academia, and research, including collaboration with Sentieon, a bioinformatics data analysis company.

Dr. Jun Ye, CEO of Sentieon, expressed his expectation for the launch of Complete Genomics’ full-line sequencing products in the US market as a representative of their partner: “We look forward to expanding diverse international laboratory partnerships in the future and fully meeting the diverse needs of the life science industry.”

In October 2022, Sentieon and MGI jointly released the DNAscope WGS model adapted to MGI sequencing platforms, which has been widely praised by users since its release. Building on this success, the Sentieon team released the WES model adapted to the MGI platform in January 2023, significantly improving the accuracy of the WES analysis process on the MGI platform.

The basic principle of Sentieon DNAscope process inherits the haplotype assembly used by GATK and improves the active region detection and local de-novo assembly steps, which enhances the sensitivity and stability of variant detection. The DNAscope process generates many potential variant sites, provides rich quality annotation information for each site, and then uses corresponding machine learning models for analysis to filter false positives and assign genotypes to true variants. The DNAscope machine learning analysis process can match different sequencing platforms and various data types generated by library preparation and capture reagents using different model files. In a recent methodological preprint, the data showed that the accuracy of DNAscope for whole-genome and whole-exome sequencing far exceeded that of standard open-source analysis processes. Sentieon will continue to release machine learning model files for different sequencing platforms for user download.

The recently launched analysis model for whole exome sequencing (WES) on the BGI platform has significantly improved the accuracy of variant detection compared to opensource software such as GATK. We summarize the key points regarding the training method and performance of this model as follows:

Training method:

  1. The training and testing samples are the standard sample HG001 (NA12878).
  2. The training and testing data features include sequencing platforms such as MGISEQ-200/500/2000 (StandardMPS, CoolMPS, and HotMPS)/T7, sequencing depths of 50x-150x, reads lengths of 100PE and 150PE, and capture kits from brands such as Agilent and MGI.
  3. The WES coverage area includes an average of 21,000 SNPs and 400 INDELs.

Model performance:

  1. The accuracy of the DNAscope WES model is significantly higher than that of GATK.
  2. The variant detection speed of the DNAscope process is 5-10 times faster than that of opensource processes.
  3. Increasing the sequencing depth and reads length can improve the accuracy of variant detection.

Based on the performance test results above, we recommend users with WES analysis needs using the MGI sequencing platform to consider upgrading their analysis workflow to DNAscope+WES model to improve both speed and accuracy. Click on the “Read the original article” link at the end of the article to access the latest model files. Existing users running the new model files need to contact Sentieon to upgrade their license for free.

To make it more convenient for users to use this integrated solution, MGI announced at the 2022 ASHG that it will provide Sentieon license for free to some users in North America, in combination with the sequencer, to form a complete variant detection solution.

Top