Hi, I'm running the differential abundance workflow with deseq_lfc_threshold parameter. As I understand it, this changes the null hypothesis used by the deseq. However, I've found that using different values for this parameter results in the same number of significant genes. Furthermore, gene lists include hits that should not come up significant under the particular lfc threshold, for instance a gene with 0.5 lfc when the threshold is set to 2. I have encountered this in both versions 1.4 and 1.5. I am including the relevant section of the params file.
Description of the bug
Hi, I'm running the differential abundance workflow with deseq_lfc_threshold parameter. As I understand it, this changes the null hypothesis used by the deseq. However, I've found that using different values for this parameter results in the same number of significant genes. Furthermore, gene lists include hits that should not come up significant under the particular lfc threshold, for instance a gene with 0.5 lfc when the threshold is set to 2. I have encountered this in both versions 1.4 and 1.5. I am including the relevant section of the params file.
Command used and terminal output
DESeq2 specific options (RNA-seq only)
deseq2_test: Wald deseq2_fit_type: local deseq2_sf_type: ratio deseq2_min_replicates_for_replace: 7 deseq2_use_t: false deseq2_independent_filtering: true deseq2_lfc_threshold: 1 deseq2_alt_hypothesis: greaterAbs deseq2_p_adjust_method: fdr deseq2_alpha: 0.1 deseq2_minmu: 0.5 deseq2_vs_method: vst deseq2_vst_nsub: 3000 deseq2_shrink_lfc: true deseq2_vs_blind: true
Relevant files
No response
System information
nexflow info: Version: 23.10.1 build 5891 Created: 12-01-2024 22:01 UTC (23:01 CEST) System: Linux 3.10.0-693.21.1.el7.x86_64 Runtime: Groovy 3.0.19 on OpenJDK 64-Bit Server VM 20.0.2-internal-adhoc..src Encoding: UTF-8 (ANSI_X3.4-1968)
Container singularity, nf-core/differentialabundance 1.5