Skip to main content
Premium Trial:

Request an Annual Quote

Swamped in Sequence Data? Consed Developer Says Autofinish Could Speed Completion


Interest in automated finishing techniques for shotgun sequence data has certainly increased in recent months, according to David Gordon, co-author of the Autofinish program that is part of the University of Washington’s Consed package.

Gordon said the program is “getting more and more known” as labs scramble to boost their finishing rates to match automated sequencing rates.

Since most finishing decisions, such as which directed reads to obtain, are still made by people, many labs have found that the gap between sequenced and finished data has widened as automated sequencers pump out data at an ever-increasing rate. Gordon and his colleagues designed Autofinish to automatically suggest the optimum finishing reads for closing the gaps in a sequencing project.

“The thing that makes Auto- finish unique is that it provides objective criteria to guide the finishing process rather than having ad hoc rules,” said Gordon. “You set a target error rate and it will finish to achieve that target error rate, trying to minimize the cost in doing so.”

Gordon said he’s not aware of any other automated software tools that use the objective criteria of error probabilities to guide the selection process.

Gordon, along with the University of Washington’s Phil Green and Cindy Desmarais, recently published a paper in Genome Research detailing a comparison between a human finisher and an Autofinish/human hybrid approach for five different sequencing projects. While the Autofinish-hybrid and the human finisher required roughly the same number and type of reads, the Autofinish-hybrid completed the five sequencing projects in 17.8 hours compared to 48.2 hours for the human finisher alone.

While labs could save a great deal of time using Autofinish, Gordon cautions that it does not yet completely automate the finishing process. He recommends that Autofinish be used for three rounds completely automatically with no human intervention. After the three rounds, a human finisher should check if there is still any further manual work that needs to be done.

As part of the Consed package, Autofinish is available for free to academic labs and may be licensed by commercial entities for $2,500 per site, regardless of the number of users or computers.

“Clearly we’re not in it for the money,” said Gordon. “The reason that we sell it at all is because that money supports further development of Consed and Autofinish.”

Consed and Autofinish are currently licensed by around 900 sites in 36 countries. Current users include the Genome Sequencing Center at Washington University, the Department of Energy Joint Genome Microbial Project, The Berkeley Drosophila Genome Project at Lawrence Berkeley National Laboratory, and a number of other private and public labs.

— BT

Filed under

The Scan

Drug Response Variants May Be Distinct in Somatic, Germline Samples

Based on variants from across 21 drug response genes, researchers in The Pharmacogenomics Journal suspect that tumor-only DNA sequences may miss drug response clues found in the germline.

Breast Cancer Risk Gene Candidates Found by Multi-Ancestry Low-Frequency Variant Analysis

Researchers narrowed in on new and known risk gene candidates with variant profiles for almost 83,500 individuals with breast cancer and 59,199 unaffected controls in Genome Medicine.

Health-Related Quality of Life Gets Boost After Microbiome-Based Treatment for Recurrent C. Diff

A secondary analysis of Phase 3 clinical trial data in JAMA Network Open suggests an investigational oral microbiome-based drug may lead to enhanced quality of life measures.

Study Follows Consequences of Early Confirmatory Trials for Accelerated Approval Indications

Time to traditional approval or withdrawal was shorter when confirmatory trials started prior to accelerated approval, though overall regulatory outcomes remained similar, a JAMA study finds.