b/tMSI-CAST (blood/tissue MSI Caller Adjusted with Sequence duplicaTes), an NGS-based composite MSI detection method that has a two-mode caller compatible with both tissue and plasma samples and does not require matched normal samples.
b/tMSI-CAST is written in Perl. Directly download the entire directory and run msicast.pl to perform MSI dectection on a single test sample. No extra compilation is required.
Version 0.10 Usage: perl msicast.pl [options]
-database <string> the directory that contains baseline databases.
-bam <string> the path of input BAM file.
-flank <int> spanning reads are defined as fully covering the coordinates of the microsatellite in reference plus flank bp in both 5’ and 3’ directions, default=2.
-num <int> the minimum allele supporting reads post-deduplication, default=2.
-minsize <int> minimum family size consdiered during deduplication, default=1.
-quality <int> minimum mapping quality, default=1.
-mode <string> available modes are 'DelRatio' and 'KLD', default=DelRatio.
-outdir <string> output directory.
-prefix <string> prefix for output files.
An example command to run MSI detection on a single pre-deduplication BAM is:
perl $0 -database <database> -bam `bam` -flank 2 -num 2 -minsize 1 -quality 1 -mode DelRatio -outdir MSI -prefix `prefix`
The program takes a pre-deduplication BAM as input, perform realignment, duplication removal, call loci-level MSI status and sample-level MSI status.
Output files include:
-
prefix.result.txt Each entry contains spanning reads information (readID, microsatellite sequence, read sequence, CIGAR and loci name) from the input BAM file. -
prefix.DUP.txt Each row contains the loci name, reference name, aligned starting position, insert size, lengths of each members pertaining to a family. -
prefix.stat.txt Each row represents info. of all fragments covering a locus, including DeletionRatios/KLD score, dup-ratios, and allele length offsets compared to the germline allele and corresponding frequencies. -
prefix.result.txt One line sample-level MSI calling results.
We can be reached by: Lili Zhao, zhao.lili@genecast.com.cn; Hongyu Xie, xie.hongyu@genecast.com.cn