Troubleshooting

Troubleshooting

loqculate: command not found

The environment that ran the install is not the one on PATH. Use uv run loqculate from the checkout, or see Installation.

--fast requires --model piecewise_cf

--fast is PiecewiseCF only. The Python loq(method="delta") path does not fall back to bootstrap; the CLI flag does, once, when delta LOQ is infinite.

Infinite LOD or LOQ

A stored result. Causes depend on the model:

• PiecewiseCF / PiecewiseWLS: $a\le 0$, too few noise observations, LOD above $\max(x)$, or the CV grid never holds <= cv_thresh for sliding_window consecutive grid points.
• Delta: $n_L<3$, no covariance, or the kink band covers the crossing.
• OriginalWLS: different LOD inner-point guard; LOQ needs a grid point with CV strictly < threshold.
• EmpiricalCV / OriginalCV: no LOD. LOQ infinite if no qualifying concentration level.

Look at --plot y before lowering cv_thresh.

--cv_thresh or --bootreps seemed to do nothing

CLI reference: several flags are parsed always and forwarded only to some models. OriginalWLS and OriginalCV compute LOQ during fit() from constructor defaults.

Peptide count looks low

Unmapped filenames are dropped. Join rules: Input formats.

WARNING: Use DIA-NN diann_report.tsv instead of pr_matrix

Header had Stripped.Sequence without Precursor.Quantity. Detection chose diann_matrix.

Generic CSV still wants a map path

The second positional is always required. Generic layout does not join on it.

Skyline or Spectronaut not detected

Header substrings, not extensions. Skyline needs Total Area Fragment, Peptide Sequence, and File Name. Spectronaut needs PEP.StrippedSequence. Pass --format if the header was rewritten.

No figuresofmerit.csv from compare

compare only writes overlay plots.

ERROR processing ... on one peptide

Logged to stderr. Row written with infinite LOD/LOQ. The run continues.