QuNex 1.5.1 release

News
1

Hi!

We are in the process of releasing QuNex 1.5.0, it will be ready within about a week or so. 1.5.0 is one of our largest updates so far and includes the following changes:

  • Made a major under the hood code refactor that should make QuNex faster and more stable.

  • Integrated the HCP Pipelines hcp_fmri_stats command into QuNex.

  • Integrated the HCP Pipelines hcp_corr_thick command into QuNex.

  • Added support for the whole HCP Transmit Bias pipeline hcp_transmit_bias_individual_align, hcp_transmit_bias_group_average_fit, hcp_transmit_bias_individual_adjustment, hcp_transmit_bias_group_average_corrected_maps, and the utility command create_transmit_bias_voltages_file to generate voltages file for phase 4.

  • Changed the QC image resolution to 2560x1080.

  • Added the parameter --clean_dicom_folders to the import_dicom command. This will perform some basic checks on the input dicom files and filter out data that is definitely not OK (e.g., empty files…).

  • Made FreeSurfer manual edits easier in the hcp_freesurfer command via the hcp_fs_edits parameter.

  • Improvements to the run_recipe framework to make it more robust and flexible.

  • Fixed some bugs in the QuNex environment setup that popped up on certain systems.

  • Mapping now supports an or operator (||), for example HiRes_T1w || LowRes_T1w => T1w will map HiRes_T1w as T1w if available and fallback to LowRes_T1w if not available.

  • Mapping now supports the asterisk operator (*), for example *_T1w will match to any label that ends with _T1w.

  • Added support for TOPUP_MISMATCH and PRECOMPUTED_FIELDMAP in hcp_fmri_volume. To be used when spin echos do not match the BOLD data in their acquisition parameters or when the fieldmap is the proper inhomogenity field (topup output).

  • Fixed a bug that sometimes caused hcp_diffusion to crash when the data had only one dir.

  • Several diffusion commands now have the optional --diffusion_folder parameter, which can be used to configure the input data that will be used.

  • Fixed a bug in longitudinal hcp_fmri_surface that looked for the presence of irrelevant files before running.

  • dwi_legacy_gpu now supports SE based distortion correction.

  • Implemented a number of internal utilities for making output folder/file backups and tracking changes through time (record_snapshot, compare_snapshots, rollback_snapshot, backup_files, restore_files). These are used by the processing pipelines and are not invocable as qunex commands.

  • Fixed a set of log bookkeeping bugs: the runlog folder is no longer mis-derived for study paths containing comlogs, the session id is written to the runlog instead of the console, the final report of a parallel run again shows the log path, and serial runs no longer record a spurious “Unknown” status next to the real one for every session.

  • Added NHP (non-human primate) support to all HCP commands that now support it. This is mainly accessible through the hcp_species parameter, while there is a completely new command for HCP FreeSurfer (hcp_nhp_freesurfer).

  • Log structure rework, logs are now in the logs folder in the study and are group per each command invocation.

  • Reworked logging into a single implementation: one runlog per invocation and one comlog per command, configurable per study or user in qunex_settings.yaml and per call with --logging, --comlog_folders, --keep_comlogs and --runlog_content; --log now sets comlog retention only, and a run in which any session fails exits non-zero.

  • QuNex study structure is now simplified, only a couple of core folders will be created initially, others are added as needed.

  • Reimplemented import_dicom to remove interim steps, set aside orphaned and non-image files, and write detailed per session report.

  • Seven utility commands (split_fidl, merge_sessions_list, import_hcp, import_nhp, map_bids2nii, run_nil_folder, bruker_to_dicom) reported a failure by printing it and still exited 0; they now report it through the log and exit non-zero, with sessions after the failing one still processed.

  • create_stats_report and check_fidl are now implemented in Python with matplotlib figures instead of calling R scripts, which also fixes check_fidl --fidlfile and --plotfile, never having worked, and the --mov_fidl/--mov_post value fd, which should have read mov.

  • Added run_qc_summary, a study level command that compiles the quality control information the processing pipelines already wrote into study tables and a self-contained interactive HTML report.

Once it is released, please report all issues you found via this forum. Thanks!

Best, Jure

2

All tests worked, so QuNex 1.5.0 is officially out! You can find it at usual locations (Docker hub or https://pooh.mblab.si/jd/qunex/qunex_suite-1.5.0.sif) under the 1.5.0 tag.

Note that we are working hard to also provide FreeSurfer 8 support for pipelines into QuNex. We planned this to go out with 1.5.0, but there are some minor hicups to resolve, expect this in the 1.5.1 version. FreeSurfer 8 container versions will be released with an _fs8 suffix. So 1.5.1_fs8. I will let you know once that is ready.

Best, Jure

3

Hi,

QuNex 1.5.1 will be released today or tomorrow. It fixes issues 1.5.0 had with the FreeSurfer8 version of the container. There are now 2 containers released the one versioned as 1.5.1 uses FreeSurfer 6 like all previous versions, the one tagged as 1.5.1_fs8 uses FreeSurfer 8. As always, let us know if there are any issues. Thanks!

Best, Jure