Li Song

GEO files usually only contain gene expression matrix and have no sequence (nucleotides) information. TRUST4 needs the raw sequence information either in the raw fastq file or aligned BAM file...

I think the easiest way is to "grep -v out_of_frame trust_report.tsv > filtered_report.tsv" first and then run VDJTools. I will check how VDJTools handle those out-of-frame recombinations.

By the way, we have a script "trust-stats.py" in the repo that can calculate commonly used diversity statistics.

One way for that is to sum up the first column (count) in the report file. Another way is to use the results from trust-stats.py, and use the formula 1000*Richness...

How did you generate the table including those tcrLibSize? We can traceback how those numbers were generated. For the CDR3 length, if you apply standardization, the mean length would be...

I think the library size is the total number of reads falling into the IGH/TRB/... region. The count in TRUST4 is the reads from the CDR3 region. So both numbers...

I think tcrCount/tcrLibsize * 1000 should be the value of CPK. But I'm not sure in the table, whether the tcrCount means the number of unique TCRs or the number...

I think you are right, the last column of TRUST3's txt output is the number of reads for the contigs containing CDR3. So this is kind of the CDR3 read...

There is a gap in the 3' part of V gene, so it could not provide the information on CDR1, CDR2 and part of CDR3. This assembly also created a...

Is your input data to TRUST4 from BAM file or raw fastq file? It is also possible that the immune infiltration level in the sample is too low to find...