v1.0.12

v1.0.12 (2024-10-24)

Main features

• Improved the predict submodule by implementing multiple attempts to load the Splam model, improving robustness and reliability.
• Introduced model/splam_static.pt, which now stores the model in state_dict format for flexibility in loading pre-trained weights.

Documentation:

The link to the documentation: http://ccb.jhu.edu/splam/

v1.0.11

v1.0.11 (2024-09-27)

Main features

• Fixed Splam GFF file parsing error when ':' is present in the gene or transcript ID.
• Added the OMP_NUM_THREADS=1 flag in setup.py to disable OpenMP multi-threading.

Documentation:

The link to the documentation: http://ccb.jhu.edu/splam/

v1.0.10

v1.0.10 (2023-11-16)

Main features

• Update Spam extraction mode to support both (1) gene - transcript - exon and (2) transcript - exon hierarchies.
• Add new test example at: https://github.com/Kuanhao-Chao/splam/blob/main/test/script_annotation_lncRNA.sh

Documentation:

The link to the documentation: http://ccb.jhu.edu/splam/

v1.0.9

v1.0.9 (2023-09-20)

Main features

• Fix the function for extracting chromosome sizes from the FASTA file
• Add two arguments for the Splam extract mode: --fr and --rf. These arguments set the strand if the alignment does not have an XS tag.
  • --fr: assume stranded library fr-secondstrand
  • --rf: assume stranded library rf-firststrand

Documentation:

The link to the documentation: http://ccb.jhu.edu/splam/

v1.0.8

v1.0.8 (2023-09-20)

Main features

• Fix PyPi README rendering issue

Documentation:

The link to the documentation: http://ccb.jhu.edu/splam/

v1.0.3 release

v1.0.3 (2023-08-28)

Main features

• Remove -A / --assembly-report argument. Splam now directly reads the length of each chromosome through fasta file and can work with any assemblies.

Documentation:

The link to the documentation: http://ccb.jhu.edu/splam/

v1.0.2 initial release

v1.0.2 (2023-07-31)

This is the initial release of the Splam code and documentation.

Documentation:

The link to the documentation: http://ccb.jhu.edu/splam/

Main features

• Trained deep residual convolutional neural network Splam model - Pytorch and Torchscript
• Extracting splice junctions in alignment files or introns in annotation files splam extract - doc link
• Scoring extracted splice junctions or introns using Splam model splam score - doc link
• Cleaning up spurious splice junctions in alignment files splam clean - doc link

Scripts for model training & testing

• The scripts for Splam model training, testing, and data analysis are available at this GitHub repository

PyPi release

• Splam v1.0.2 is also available on PyPi: https://pypi.org/manage/project/splam/release/1.0.2/