diann-1.8.1 different results between linux and windows

[ydzdfci]

Environment

• DiaNN version: 1.81.
• OS 1: AlmaLinux release 8.10
• Hardware 1: AMD Ryzen Threadripper PRO 5965WX 24-Cores with 256G RAM
• OS 2: Windows 11
• Hardware 2: Intel i9-10900k with 32G RAM

Log

-Linux

DIA-NN 1.8.1 (Data-Independent Acquisition by Neural Networks)
Compiled on Apr 15 2022 08:45:18
Current date and time:
Logical CPU cores: 48
./diann-1.8.1 --f file1.d --f file2.d --f file11.d --lib  --threads 40 --verbose 1 --out report.tsv --qvalue 0.01 --matrices --out-lib report-lib.tsv --gen-spec-lib --predictor --fasta sample.fasta --fasta-search --min-fr-mz 200 --max-fr-mz 1800 --met-excision --cut K*,R* --missed-cleavages 2 --min-pep-len 7 --max-pep-len 30 --min-pr-mz 300 --max-pr-mz 1800 --min-pr-charge 2 --max-pr-charge 4 --unimod4 --var-mods 2 --var-mod UniMod:35,15.994915,M --reanalyse --relaxed-prot-inf --smart-profiling --peak-center 

Thread number set to 40
Output will be filtered at 0.01 FDR
Precursor/protein x samples expression level matrices will be saved along with the main report
A spectral library will be generated
Deep learning will be used to generate a new in silico spectral library from peptides provided
Library-free search enabled
Min fragment m/z set to 200
Max fragment m/z set to 1800
N-terminal methionine excision enabled
In silico digest will involve cuts at K*,R*
Maximum number of missed cleavages set to 2
Min peptide length set to 7
Max peptide length set to 30
Min precursor m/z set to 300
Max precursor m/z set to 1800
Min precursor charge set to 2
Max precursor charge set to 4
Cysteine carbamidomethylation enabled as a fixed modification
Maximum number of variable modifications set to 2
Modification UniMod:35 with mass delta 15.9949 at M will be considered as variable
A spectral library will be created from the DIA runs and used to reanalyse them; .quant files will only be saved to disk during the first step
Highly heuristic protein grouping will be used, to reduce the number of protein groups obtained; this mode is recommended for benchmarking protein ID numbers; use with caution for anything else
When generating a spectral library, in silico predicted spectra will be retained if deemed more reliable than experimental ones
Fixed-width center of each elution peak will be used for quantification
DIA-NN will optimise the mass accuracy automatically using the first run in the experiment. This is useful primarily for quick initial analyses, when it is not yet known which mass accuracy setting works best for a particular acquisition scheme.
Exclusion of fragments shared between heavy and light peptides from quantification is not supported in FASTA digest mode - disabled; to enable, generate an in silico predicted spectral library and analyse with this library

11 files will be processed
[0:00] Loading FASTA sample.fasta
WARNING: 38 sequences skipped due to duplicate protein ids; use --duplicate-proteins to disable skipping duplicates
[0:06] Processing FASTA
[0:14] Assembling elution groups
[0:22] 7318876 precursors generated
[0:22] Gene names missing for some isoforms
[0:22] Library contains 5730 proteins, and 5730 genes
[0:24] [0:33] [18:04] [19:42] [19:49] [19:51] Saving the library to report-lib.predicted.speclib
[20:02] Initialising library

-Windows

DIA-NN 1.8.1 (Data-Independent Acquisition by Neural Networks)
Compiled on Apr 14 2022 15:31:19
Current date and time: 
CPU: GenuineIntel Intel(R) Core(TM) i9-10900K CPU @ 3.70GHz
SIMD instructions: AVX AVX2 FMA SSE4.1 SSE4.2 
Logical CPU cores: 20
diann.exe --f file1.d  --f file2.d  --f file11.d  --lib  --threads 20 --verbose 1 --out report.tsv --qvalue 0.01 --matrices --out-lib report-lib.tsv --gen-spec-lib --predictor --fasta sample.fasta --fasta-search --min-fr-mz 200 --max-fr-mz 1800 --met-excision --cut K*,R* --missed-cleavages 2 --min-pep-len 7 --max-pep-len 30 --min-pr-mz 300 --max-pr-mz 1800 --min-pr-charge 2 --max-pr-charge 4 --unimod4 --var-mods 2 --var-mod UniMod:35,15.994915,M --reanalyse --relaxed-prot-inf --smart-profiling --peak-center 

Thread number set to 20
Output will be filtered at 0.01 FDR
Precursor/protein x samples expression level matrices will be saved along with the main report
A spectral library will be generated
Deep learning will be used to generate a new in silico spectral library from peptides provided
Library-free search enabled
Min fragment m/z set to 200
Max fragment m/z set to 1800
N-terminal methionine excision enabled
In silico digest will involve cuts at K*,R*
Maximum number of missed cleavages set to 2
Min peptide length set to 7
Max peptide length set to 30
Min precursor m/z set to 300
Max precursor m/z set to 1800
Min precursor charge set to 2
Max precursor charge set to 4
Cysteine carbamidomethylation enabled as a fixed modification
Maximum number of variable modifications set to 2
Modification UniMod:35 with mass delta 15.9949 at M will be considered as variable
A spectral library will be created from the DIA runs and used to reanalyse them; .quant files will only be saved to disk during the first step
Highly heuristic protein grouping will be used, to reduce the number of protein groups obtained; this mode is recommended for benchmarking protein ID numbers; use with caution for anything else
When generating a spectral library, in silico predicted spectra will be retained if deemed more reliable than experimental ones
Fixed-width center of each elution peak will be used for quantification
DIA-NN will optimise the mass accuracy automatically using the first run in the experiment. This is useful primarily for quick initial analyses, when it is not yet known which mass accuracy setting works best for a particular acquisition scheme.
Exclusion of fragments shared between heavy and light peptides from quantification is not supported in FASTA digest mode - disabled; to enable, generate an in silico predicted spectral library and analyse with this library

11 files will be processed
[0:00] Loading FASTA sample.fasta
WARNING: 38 sequences skipped due to duplicate protein ids; use --duplicate-proteins to disable skipping duplicates
[0:09] Processing FASTA
[0:23] Assembling elution groups
[0:38] 8645179 precursors generated
[0:38] Gene names missing for some isoforms
[0:38] Library contains 5733 proteins, and 5733 genes
[0:38] [0:50] [55:15] [68:22] [68:35] [68:41] Saving the library to report-lib.predicted.speclib
[68:51] Initialising library

From the log above, based on the same fasta file, linux has 7318876 precursors generated while windows has 8645179 precursors generated. This will cause different results.

[vdemichev]

Hi,

Thank you for reporting this, apparently a bug in the Linux version of 1.8.1. In general, please use 1.9.2, should not have any differences between the platforms.

Best,
Vadim

[ydzdfci]

Thanks Vadim.
Our lab is planning to use 1.9.2. For now, I tried to use 1.9.2, but it seems that some parameters are different from version 1.8.1. Like no "--reanalyse"
Do you have command conversion tool or do you know what's the equivalent command as below in 1.9.2? Than
./diann-1.8.1 --f file1.d --f file2.d --f file11.d --lib --threads 40 --verbose 1 --out report.tsv --qvalue 0.01 --matrices --out-lib report-lib.tsv --gen-spec-lib --predictor --fasta sample.fasta --fasta-search --min-fr-mz 200 --max-fr-mz 1800 --met-excision --cut K*,R* --missed-cleavages 2 --min-pep-len 7 --max-pep-len 30 --min-pr-mz 300 --max-pr-mz 1800 --min-pr-charge 2 --max-pr-charge 4 --unimod4 --var-mods 2 --var-mod UniMod:35,15.994915,M --reanalyse --relaxed-prot-inf --smart-profiling --peak-center
I will compare the results difference again between different version and different OS
Thanks for your time.

[vdemichev]

Do you have command conversion tool or do you know what's the equivalent command as below in 1.9.2?

Please see the command generated by DIA-NN GUI. You can also do this on Linux under Wine.