I’m using bracer to process my single cell data but I found that the assembled contigs are not so similar to the IMGT reference according to the result of Igblast.
For example, when I used Tracer to porcess the data, the similarity (identity(%) by Igblast) between the assembled contigs and the IMGT V/J reference are almost more than 99%, like the example below.
When I used Bracer, the similarity seemed to be very low, often less than 95%,even less than 90% in some samples, like the examples below.
As a result , most of (more than 90%) T cells or Macrophages in my data also had a positive result(bearing BCR) according to the bracer output.
So I‘m wondering if the “synthetic BCR genomes” was designed well enough to extract BCR-derived reads. It seems that many false positive reads were included, effecting the down stream analysis results.
I’m using bracer to process my single cell data but I found that the assembled contigs are not so similar to the IMGT reference according to the result of Igblast.
For example, when I used Tracer to porcess the data, the similarity (identity(%) by Igblast) between the assembled contigs and the IMGT V/J reference are almost more than 99%, like the example below.![image](https://user-images.githubusercontent.com/42466569/53940053-a02a5100-40ef-11e9-9e1e-5ba746aabe43.png)
When I used Bracer, the similarity seemed to be very low, often less than 95%,even less than 90% in some samples, like the examples below.![image](https://user-images.githubusercontent.com/42466569/53940292-45452980-40f0-11e9-87c3-8bf49f22c8e8.png)
As a result , most of (more than 90%) T cells or Macrophages in my data also had a positive result(bearing BCR) according to the bracer output.
So I‘m wondering if the “synthetic BCR genomes” was designed well enough to extract BCR-derived reads. It seems that many false positive reads were included, effecting the down stream analysis results.