How does the machine know a read belongs to a certain index?

[szy0931]

I have a very basic question.
How does the machine know a read belongs to a certain index?
let's say we have a sequence like this in a single end RNA-Seq:
5' end adapter-cDNA-index-3' end adapter
For each read, sequencing starts from 5' end and ends a certain place with cDNA because the fragment of 5' end adapter-cDNA-index-3' end adapter is longer than the reads.
Apparently, the index is not contained the reads.
So how does the machine know a read belongs to a certain index?

[GenoMax]

Machine (actually the basecalling software) knows that a particular index read belongs to a certain cluster the read came from (software tracks physical position X,Y coordinates of that cluster in a lane, that information is in the read header).

[szy0931]

GenoMax,

Thank you!

[GW_OK]

In the same way that the machine knows that each cluster of light forms the bases for each read: the pixels encompassing each light signal are precisely mapped. When the machine "finds" the clusters, it is actually mapping which pixels are giving signal.

[szy0931]

So, how does the machine or software know which index is bound to it?

[GenoMax]

So, how does the machine or software know which index is bound to it?

It doesn't "know" which index (demultiplexing software takes care of that assignment) since indexes are just another stretch of nucleotides.

A particular physical cluster (identified/tracked by the software) generates a read (R1/R2). During the index read sequencing (which happens independently in Illumina) the sequence coming from a specific cluster (identified solely by its physical location, tracked in header ID) allows the software to associate the index sequence to that cluster (and ultimately to the actual reads that came from that cluster).

[szy0931]

So, there are two sequencing events happen on a cluster. One is to sequence the inner cDNA. The other one is to sequence (read) the index to tell the position of the first sequencing event.
Is this right?

[GenoMax]

So, there are two sequencing events happen on a cluster. One is to sequence the inner cDNA. The other one is to sequence (read) the index to tell the position of the first sequencing event.
Is this right?

Up to 4 events can happen (if it was a PE 2D run).

Event 1 --> R1
Event 2 --> Index 1 (optional)
Event 3 --> Index 2 (optional)
Event 4 --> R2 (optional)

Watch this video for a visual explanation.