genomics automation
Photo courtesy of Indra Projects
 
 Short-read sequencing built the modern genomics era. Illumina platforms, generating millions of short DNA fragments in parallel, became the workhorse of most laboratories from the late 2000s onward. Now, according to genomics specialist Manan Chandra, long-read sequencing is starting to reshape what those laboratories can ask in the first place.
 
Chandra has worked across both formats, with roughly eleven years in the sequencing industry covering Illumina short-read platforms, Oxford Nanopore long-read sequencing, and earlier Sanger sequencing work.
 
Three Areas Where Long-Read Is Changing The Question
Long-read platforms read continuous stretches of DNA tens or hundreds of times longer than a typical short-read fragment. That difference matters most in three places, Chandra says: genome assembly, where long reads can span repetitive regions that short reads cannot resolve; structural variant analysis, where long reads can detect larger insertions, deletions, and rearrangements that short reads tend to miss; and microbial genomics, where long reads help reconstruct complete bacterial chromosomes and plasmids from mixed samples.
 
"I'm especially interested in how next-generation sequencing continues to evolve through areas such as long-read sequencing technologies like Oxford Nanopore, which are expanding possibilities in genome assembly, structural variant analysis, and microbial genomics," he said.
 
None of those capabilities is entirely new. The shift, Chandra suggests, is that they are starting to become routine.What Changes Inside The Laboratory
The harder part of the transition, in his view, sits below the headline capabilities. Long-read workflows introduce new sample-quality requirements, particularly around DNA fragment length and integrity. Library preparation looks different. So does quality control. So does data interpretation, where laboratories accustomed to short-read pipelines have to rebuild parts of their analysis stack.

Chandra has worked extensively on the operational side of sequencing, including protocols for low-input and lower-quality DNA and RNA samples, normalization processes designed for higher-throughput environments, and dilution strategies for enzyme-based reactions suited to semi-automated processing. He treats those steps as the place where most of a sequencing run's reliability is actually determined, before any read ever reaches a flow cell.

His research record runs alongside the applied work, across transcriptomics, microbial genomics, and gene regulation. His most cited contribution studied alternative splicing responses in pathogen-infected Arabidopsis plants, using RNA-Seq and transcript-isoform-specific validation to track how the plants regulate defense during infection; the resulting PLOS ONE paper has been cited roughly 150 times. He has also contributed to a review publication on Virus-Induced Gene Silencing in plants, and is now preparing research on sulfur metabolism and biofilm regulation in Staphylococcus aureus, looking at how metabolic pathways shape bacterial adaptation and gene expression. He earned a Master's degree, with thesis, in Functional Genomics from North Carolina State University.

Where Automation And AI Fit In
Chandra is careful not to overstate the role of machine learning in the near term, but he does see it converging with the long-read shift. Automation makes sequencing workflows more reproducible at scale. AI-driven analysis, he argues, is becoming a more practical tool for quality control and for interpreting datasets that are now too large and too complex for manual review at every step.

Long-read sequencing produces a different kind of data problem than short-read sequencing does. More signal, more nuance, more places where interpretation can go wrong. Laboratories that solve the operational and the analytical sides together, Chandra suggests, will pull ahead. Those who treat the new platforms as a drop-in replacement for the old ones tend to discover, slowly, that they are not.