Repository navigation
DIANN
Conduit relies on DIA-NN for spectral matching and quantification.
DIA-NN performs the core computational tasks of matching MS/MS spectra to peptides and estimating peptide abundances.
For background, see the original DIA-NN publication:
Demichev et al., Nature Methods (2019) — https://www.nature.com/articles/s41592-019-0638-x
DIA-NN is integrated into Conduit in two main steps:
- Generating a spectral library
- Running the DIA-NN search
This step involves predicting retention times and spectral properties for in silico–digested peptides.
DIA-NN internally uses deep learning–based spectral predictors Prosit- sytle models to estimate fragment ion intensities and retention times.
Conduit allows users to customize this step using any flags or parameters supported by DIA-NN’s command-line interface:
👉 DIA-NN Command-Line Reference
By default:
- The file
generate_diann_spectral_library.cfgfrom the top-level configuration directory is copied to your experiment-specific config folder. - This default file contains sensible baseline parameters suitable for most experiments.
For fine-grained control, you can provide your own generate_diann_spectral_library.cfg in your experiment’s config directory.
Conduit will automatically use your version instead of the default.
Once the spectral library is generated, DIA-NN is run to search the experimental MS data and identify peptide-spectrum matches (PSMs).
Similar to the spectral library generation step, Conduit gives users full control over the search parameters — any command-line flag accepted by DIA-NN can be used.
By default:
- The file
run_diann.cfgfrom the top-level configuration directory is copied into your experiment-specific config directory. - This file includes default parameters designed for robust and reproducible results.
For more detailed customization, you can supply your own run_diann.cfg file with arbitrary values.
Conduit will use your configuration to control how DIA-NN performs the search.
In summary:
Conduit provides robust defaults for both DIA-NN spectral library generation and searching, while allowing full parameter transparency and user control for advanced workflows.