search

xjtu-omics / msisensor-pro

Microsatellite Instability (MSI) detection using high-throughput sequencing data.
Other
98 stars 22 forks source link

MSI score greater than 1 #11

Closed shanshenbing closed 4 years ago

shanshenbing commented 4 years ago

I ran MSIsensor-pro with paired samples successfully and I found some sample has a MSI score great than 1. Actually, the MSI score is the percent of somatic sites, which can not be 1 or great than 1. So, I am confused with the result. I hope you can give me some suggestions. 25X(X8GVQDO$8K8TFKZBNH0

PengJia6 commented 4 years ago

I ran MSIsensor-pro with paired samples successfully and I found some sample has a MSI score great than 1. Actually, the MSI score is the percent of somatic sites, which can not be 1 or great than 1. So, I am confused with the result. I hope you can give me some suggestions. 25X(X8GVQDO$8K8TFKZBNH0

Could you please send your command and output to me. My email is pengjia@stu.xjtu.edu.cn

Thanks !

shanshenbing commented 4 years ago

Sorry，It‘s my fault. I didn't notice that MSIsensor-pro output is in percent and my data is just greater than 1%. And I also run msisensor with my data and I found the MSI scores of the same sample are different. Is this normal?

PengJia6 commented 4 years ago

Sorry，It‘s my fault. I didn't notice that MSIsensor-pro output is in percent and my data is just greater than 1%. And I also run msisensor with my data and I found the MSI scores of the same sample are different. Is this normal?

This is normal, but I don't think it matters. This may caused by the different default parameters between these two programs.

shanshenbing commented 4 years ago

Thank you for your reply. The reason why I noticed the problem is that the difference is a little big.  I ran msisensor-pro scan with fasta file and using the generated file for next step.  The softwares werw run as follow:

nohup msisensor-pro msi -z 1  -b 3 -0 1 -y 1    -d /path_to/Homo_sapiens.GRCh37.75.format.list \     -t /path_to/${tumor}.reorder.sort.add.mkdup.BQSR.bam \     -n /path_to/${normal}.reorder.sort.add.mkdup.BQSR.bam  \     -o ${WORK_DIR}/result/${sample}/${sample}  &

nohup msisensor msi -z 1  -b 3  -y 1    -d /path_to/Homo_sapiens.GRCh37.75.format.list \     -t /path_to/${tumor}.reorder.sort.add.mkdup.BQSR.bam \     -n /path_to/${normal}.reorder.sort.add.mkdup.BQSR.bam  \     -o ${WORK_DIR}/result/${sample}/${sample} &

note: I used the same list file generated by msisensor-pro scan.

The result of first 35 samples is in the accessory, and I found the Pearson r is about 0.7 compulated with all my data.  The parameters are similar to each other, but results are different. And I also found a relate issue at https://github.com/ding-lab/msisensor/issues/3. 

I hope to get your advice. ------------------ 原始邮件 ------------------ 发件人: "xjtu-omics/msisensor-pro" <notifications@github.com>; 发送时间: 2020年8月11日(星期二) 晚上10:24 收件人: "xjtu-omics/msisensor-pro"<msisensor-pro@noreply.github.com>; 抄送: "单申兵"<1186426740@qq.com>;"Author"<author@noreply.github.com>; 主题: Re: [xjtu-omics/msisensor-pro] MSI score greater than 1 (#11)

This is normal, but I don't think it matters. This may caused by the different default parameters between these two program.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

PengJia6 commented 4 years ago

The result of first 35 samples is in the accessory, and I found the Pearson r is about 0.7 compulated with all my data.  The parameters are similar to each other, but results are different. And I also found a relate issue at ding-lab/msisensor#3. 

The msi module in msisensor-pro is inherited from msisensor, the results should be same. Could you please add -c option to the two command and try again, because there is a small bug in the default parameter, please see #8 .
Please let me know if this is workable and feel free to contact me if you have any question.

shanshenbing commented 4 years ago

I wrote the last comment in QQ mailbox and the space character was converted into '  '. I am sorry for the confusing meaning. I checked msisensor default parameters and found the default -c in msisensor is 20, so msisensor-pro was run with -c 20. The result of the two software are same. Thank you for your help! Maybe the next release of msisensor-pro can change the default value of -c parameter.

PengJia6 commented 4 years ago

@shanshenbing Thanks, I had updated source code and binary version, and conda version would be updated later. Thanks for using our software and give feedback, if you have other questions, please open a new issue or contact with me directly.