search

huanglikun / BRM

Block Regression Mapping (BRM) is a statistical method for QTL mapping based on bulked segregant analysis by deep sequencing
GNU General Public License v3.0
7 stars 3 forks source link

NAs produced by integer overflow #5

Closed ColyFu closed 2 years ago

ColyFu commented 2 years ago

Dear Dr. Huang,

We used BRM for BSA analysis of rice and encountered the attached error. Could you please help find the cause of the error? I can send you the input data if needed. Thank you very much!

Jun

Read global parameters from BRM_conf.txt . Parameter Design=A Parameter n1=35 Parameter n2=35 Parameter t=1 Parameter ua=3.65 Parameter UNIT=1000 Parameter DEG=2 Parameter BLK=0.2 Parameter MIN=10 Parameter MINVALID=10 Parameter Result1_File=result/result1.xls Parameter Result2_File=result/result2.xls

Number of missing data: 25697 Number of imported chromosomes is 12 . Data size of 1 is 89153 . Total blocks in 1: 216356 Number of valid block: 37493 Error in double(N M1) : vector size cannot be NA Calls: run_step1 -> predict -> predict.loess -> predLoess -> double In addition: Warning message: In N M1 : NAs produced by integer overflow Execution halted

huanglikun commented 2 years ago

Thank you for using BRM! I'll try to fix the bug and provide more warnings instead of errors later. Before the bug was fixed, you could try this approach as follow: The block size setting for rice is recommended to be 5K to 20K, meaning BLK=5 to BLK=20. Maybe BLK=0.2 for rice is too small. The suitable block size is about a physical distance of 0.1cM. P.S. BLK=0.2 is a recommended setting for yeast.

ColyFu commented 2 years ago

It worked like a charm. Thank you for your prompt reply.