Output problem
Opened this issue · 12 comments
Hi, I have some questions while using the msisensor2，I specify 27 sites for input, and only 7 sites have enough coverage (show below):
But, the distribution of read counts as below:
So, I don't understand what happened, can you help me, thanks!
Have you ever set the -c parameter ? Can you show me your whole command ?
Have you ever set the -c parameter ? Can you show me your whole command ?
@Beifang Hi, Beifang.
Command line ：
msisensor2 msi -M models_hg19/ -d 27_marker_microsatellites.list -t test1.bam -p 5 -o msi
I have set the -c from 0 to 20, but the results are the same.
By using machine learning, only hypersensitie microsatellite sites were singled out for MSI detection. So, there are only 7 sites overlap between 27 sites you specified and msisensor2 model set. By our robustness test, the msisensor2 algorithm is effective with >= 5 random sites for WES data with >=0.99 AUC. So, 7 sites should works for you. We will change the output message to prevent misunderstanding in msisensor2.
@Beifang, thank you very much for your detailed explanation!
But, I have another question:
- 2 of 27 msi sites are BAT-25 and BAT-26, they belong to the PCR golden standard of 5 markers, and for this sample the PCR results of these 2 sites are unstable. But BAT-25 and BAT-26 are not in the 7 sites. If i want to use them or all 27 sites, what should i do.
I totally understand and appreciate your concerns. Several users have asked us to realease the models for 7 PCR golden standard markers ( BAT25/26, NR21/22/24/27, MONO27 ) soon. So, don't worry about that and we will add these 7 markers in hg19/38 model sets.
Yes, we can send the 27 models for your test if you provide us the list.
@Beifang
Here is the 27 sites list in hg19. Thank you very much for your patient answer.
27_marker_microsatellites.txt
@Beifang
Here is the 27 sites list in hg19. Thank you very much for your patient answer.
27_marker_microsatellites.txt
@Beifang
Here is the 27 sites list in hg19. Thank you very much for your patient answer.
27_marker_microsatellites.txt
7 PCR golden standard markers and 10 sites from your 27 site list can not be captured by common WES or target capture assays. We provide the other 17 models for you :
msisensor2 msi -M ./test/tmp/models_hg19_17sites -t xxx.hg19.bam -o output.tumor.prefix
@Beifang Thank you very much for your team's help! when other account use these new model, I found the permissions needed to write, but just have read only. Maybe you forgot to change permissions.
Thanks for your nice help again.
Could you please check your command line to make sure it is correct ? and make sure your account has permission as well.
Hi, I have similar problem like xin8you, I am trying to determine msi status in BAT25, BAT26, MonoC, NR22/24/27/21, CAT25 and D5S346 using msisensor2. Are STL 3D model files avilable for these sites (or for any of them) anywhere to download?
Thank You
@xin8you can you tell me how to calculate repeat_unit_binary, left_flank_binary, right_flank_binary? I want to use the 7 PCR golden standard markers in tumor-control model.
Thank you.