“Multiplexing” and “Demultiplexing” in next generation sequencing


I’m trying to learn about next-generation sequencing techniques (specifically sc-RNA seq) and I keep seeing the word “multiplexing” or “demultiplexing” reads/data. I’ve tried to look it up but I can’t seem to find a good explanation.

In: Biology

I’m not a biologist, but in Electronic engineering, we use multiplexing for sending multiple signals down a transmission line, and demultiplexing for splitting one transmission line into many signals at the other end. It looks like this is much the same. you take many samples, combine them into one data-set, sequence them on a high-power sequencing machine, and then un-merge them into their respective data-sets

Most NGS is done on flow cells which are separated into lanes. The data generated by the machine will be grouped by lane. Normally, this would mean you can only put one library on each lane, because if you mixed libraries on the same lane the reads would all be grouped together and you wouldn’t know which read came from which library. With current Illumina HiSeq machines you can get somewhere between 300-400 million reads per lane, which is way more than most people need for a single library and would be a waste of money.

Multiplexing is the solution to this. By attaching a different 6-nucleotide “barcode” to each library, they can then be pooled on a single lane and then de-multiplexed to separate the reads at the end. For example, your library from a liver sample might have the barcode ACTGCA and your library from a lung sample might have CCTACG. After sequencing, these barcodes can be searched for in the data to separate the liver reads from the lung reads.