Losing clusters during correction stage

[mjudd8]

After the clustering stage, the output given was "20623 gene clusters found". After taking the clusters.out file and moving through the correction stage, 8074 clusters were left in the consensi.fq file. Why this drastic decrease in clusters post-correction?

[EduEyras]

Hi,

Thanks for your question

By default, RATTLE only corrects clusters with at least 5 reads You can change this in the correct command with the parameter -r, --min-reads This step will reduce the number of transcript clusters going through to the consensus stage.

In the consensus-building stage, the transcripts built from the initial clusters may be merged if they show sufficient similarity with each other. This step basically reruns the cluster step but using the representatives from the transcript clusters. This may also reduce the number of resulting transcripts.

I hope this clarifies your question

Best

Eduardo

On Fri, 1 Mar 2024 at 06:58, mjudd8 @.***> wrote:

After the clustering stage, the output given was "20623 gene clusters found". After taking the clusters.out file and moving through the correction stage, 8074 clusters were left in the consensi.fq file. Why this drastic decrease in clusters post-correction?

— Reply to this email directly, view it on GitHub https://github.com/comprna/RATTLE/issues/52, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADCZKB5DCTMQQBWTHISG7A3YV6D6DAVCNFSM6AAAAABEAR5RL6VHI2DSMVQWIX3LMV43ASLTON2WKOZSGE3DEMBQGA3DKMQ . You are receiving this because you are subscribed to this thread.Message ID: @.***>