-
Notifications
You must be signed in to change notification settings - Fork 22
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Transcripts identified by Bambu have a CPM or count value of 0 in all samples #422
Comments
Hi, I cannot be sure without seeing the results, but my best guess is that these 125,254 annotations are reference annotations that are not expressed in your sample. Bambu does not remove any annotations as it is hard to prove a transcript is not expressed, rather than simply not detected, that is why we extend the annotations, and only add transcripts to the final annotations. We then rely on the expectation maximization algorithm to reduce the impact of the potentially unexpressed transcripts. If you are only interested in transcripts which have full length support you can filter the output to only include those that were detected by following the steps under this section of the documentation Kind Regards, |
Hi Andre, Thanks for your reply! Out of the 125,254 transcripts, I found 2 novel transcripts (BambuTx584 and BambuTx1634). The remaining transcripts are reference annotations that are not expressed in my samples. Here are my output files: Here is an example of some transcripts with CPM=0 in all samples: These 2 BambuTx also with count=0 in all samples in Yuyun |
Hi Yuyun, We discussed this internally to try and determine why you are getting these results. We came up with the following:
If we cannot diagnose it from the above steps, it will be helpful for us to have a look at your intermediate files to try and solve this for you. This would involve you sending us the read class files .rds Kind Regards, |
Hi Andre, Thanks for the detailed advice! I used 3.4.0 bambu to obtain these results. I am trying to install new version (3.4.1 release 3_18) and will send the intermediate files if I successfully install the new version.
Here is output to check if BambuTx584 and BambuTx1634 are subsets of another novel transcript (It seems that BambuTx584 and BambuTx1634 are the only novel transcript from their respective genes):
Best, |
Hi Yuyun, Thanks for sending this, this was helpful to help us rule out a few possible reasons. Did running Bambu with the latest version change the results? If not, unfortunately the reason for this eludes us so we will need to delve deeper. We have two possible explanations, one being that the second example, the counts are being assigned to the other gene which overlaps these same exons, and the other being that the final transcript models are slightly different from the read classes when multiple samples are merged, so the original read class is assigned elsewhere. To solve this, are you able to rerun your analysis with Kind Regards, |
Hi Andre, Sorry for the delayed response, and thank you for providing me with advice! I installed the bambu 3.4.1 (release_3_18), but I can't use the parameter
Now I am runing the command below, and waiting for the results.
Thank you again! |
After running the command, I got an error, and I'm currently trying to resolve it. I am wondering if you have any suggestions? Thank you!
|
Hi Yuyun, Sorry for the delayed response. Thanks, |
Hi Andre, Thank you! Here is the sessionInfo() output:
|
Hi Yuyun, Very sorry I made a mistake earlier. We actually renamed assignDist to returnDistTable, and I forgot about that change. Could you please try rerun it with that instead. Regarding the other issue, this seems to be stemming from rcOutDir and I am getting the same error on my side too which I am trying to resolve now. Seems like the parameters for the bfcnew function are different now. While I fix this could you please run it as follows.
Kind Regards, |
The error Error in (function (cond) : error in evaluating the argument 'x' in selecting a method for function 'bfcnew': Failed to collect lazy table. Is due to an incompatibility between the latest dbplyr 2.5.0 and BiocFileCache. |
Hi Andre, Thanks for your information. I downgraded dbplur to 2.3.4 and the issue was fixed. After using bambu 3.4.1 (release_3_18), all novel transcripts are expressed (count value > 0 and CPM > 0) in at least 1 sample. I got the results by running following command:
I am wondering if you still need to check the rcfile.rds and se.rds file? If so, I can send these files to your email. Thank you again! Yuyun |
I am glad the issue is fixed. No need to send these files if the issue was due to version differences because of the dependencies. |
Hi,
Thanks for developing bambu, it's very helpful in my project. I have a question about bambu: why transcripts identified by Bambu have a CPM or count value of 0 in all samples?
I have 9 samples, and run all samples together. Bambu output total 210,028 transcripts, but I found there are 125,254 transcripts with CPM of 0 in all sample in the output file
CPM_transcript.txt
. And 125,003 transcripts with count value of 0 incounts_transcript.txt
.I think if a transcript can be detected and output to the
extended_annotations.gtf
, it should has support reads and the expression value may not be 0. So I am wondering if there are any rules in quantifying transcripts and correcting the value to 0?Thank you!
The text was updated successfully, but these errors were encountered: