Thanks to advancements in DNA and RNA sequencing, biologists are good at knowing how much of a gene's code is at any moment being copied into RNA messages. But they're not so good at figuring out how quickly those RNA messages are actually read from end to end at ribosomes.
Now, a multidisciplinary team of researchers from Cold Spring Harbor Laboratory (CSHL), Stony Brook University (SBU), and Johns Hopkins University (JHU) has released software that can help biologists more accurately determine this. They used single-celled yeast and E. coli to demonstrate their new program, called Scikit-Ribo. Additional details can be found in the recent Cell Systems paper.
Scikit-Ribo was designed to improve upon Riboseq, which was developed almost 10 years ago. The latter revealed which, and how quickly, cells translate RNA into protein. It was a great advance, says Michael Schatz, a quantitative biologist at CSHL and JHU, who with Gholson Lyon of CSHL, supervised the work of Han Fang, a recent graduate of SBU. It was Fang who figured out how to improve Riboseq so data could be brought into focus.
Fang's insight was to use advanced statistical modeling techniques to account for the fact that ribosomes do not work at a uniform rate, but rather tend to pause—for instance, when they encounter hairpin-shaped kinks in incoming RNA messages. Scikit-ribo also filters out noise that muddied raw Riboseq results. Now the two methods can be used together, to generate a much more accurate picture of which RNA messages are being read at specific ribosomes, and, perhaps most important, how much functional protein is being generated.
"The amount of protein that's actually made may or may not be the same as the amount that a given gene is being expressed," says Schatz. Having a more reliable way of knowing will help in disease research.