Note: I haven't yet updated to R 4.0 or Bioconductor 3.11 yet in case that matters.

I built a STAR index using GRCm38 ensembl release 98 dna primary assembly and the release 98 gtf, and aligned with the --quantMode TranscriptomeSAM to get outputs in transcriptome coordinates. I then used gffread -w transcripts.fa -g $DNA $GTF to make a reference for salmon's alignment mode (salmon quant -t transcripts.fa -l A -a Sample.toTranscriptome.out.bam -o $OUTDIR) which seems to have worked.

Unfortunately, despite using unchanged references from ensembl, I can import it with tximeta, but it does not recognize the transcriptome and only provides a non-ranged SummarizedExperiment.

In the tximeta vignette sections about creating linked transcriptomes, everything seems to require a salmon index unless I'm missing something.

Can someone help me find a way that will allow use of the automatic metadata gathering feature of tximeta?

> library(tximeta)
> meta <- tibble::tribble(
+   ~names,        ~files,
+   "sample1", "results/quant/salmon/STAR/quant.sf"
+ )
> se <- tximeta(meta)
importing quantifications
reading in files with read_tsv
1 
couldn't find matching transcriptome, returning non-ranged SummarizedExperiment

.

> sessionInfo()
R version 3.6.3 (2020-02-29)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 18363)

Matrix products: default

locale:
[1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United States.1252    LC_MONETARY=English_United States.1252 LC_NUMERIC=C                          
[5] LC_TIME=English_United States.1252    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] tximeta_1.4.5

loaded via a namespace (and not attached):
 [1] Rcpp_1.0.4.6                lattice_0.20-41             prettyunits_1.1.1           Rsamtools_2.2.3             Biostrings_2.54.0          
 [6] assertthat_0.2.1            digest_0.6.25               BiocFileCache_1.10.2        R6_2.4.1                    GenomeInfoDb_1.22.1        
[11] stats4_3.6.3                RSQLite_2.2.0               httr_1.4.1                  pillar_1.4.4                zlibbioc_1.32.0            
[16] rlang_0.4.6                 GenomicFeatures_1.38.2      progress_1.2.2              lazyeval_0.2.2              curl_4.3                   
[21] rstudioapi_0.11             blob_1.2.1                  S4Vectors_0.24.4            Matrix_1.2-18               BiocParallel_1.20.1        
[26] readr_1.3.1                 stringr_1.4.0               ProtGenerics_1.18.0         RCurl_1.98-1.2              bit_1.1-15.2               
[31] biomaRt_2.42.1              DelayedArray_0.12.3         compiler_3.6.3              rtracklayer_1.46.0          pkgconfig_2.0.3            
[36] askpass_1.1                 BiocGenerics_0.32.0         tximport_1.14.2             openssl_1.4.1               tidyselect_1.0.0           
[41] SummarizedExperiment_1.16.1 tibble_3.0.1                GenomeInfoDbData_1.2.2      IRanges_2.20.2              matrixStats_0.56.0         
[46] XML_3.99-0.3                fansi_0.4.1                 crayon_1.3.4                dplyr_0.8.5                 dbplyr_1.4.3               
[51] GenomicAlignments_1.22.1    bitops_1.0-6                rappdirs_0.3.1              grid_3.6.3                  jsonlite_1.6.1             
[56] lifecycle_0.2.0             DBI_1.1.0                   AnnotationFilter_1.10.0     magrittr_1.5                cli_2.0.2                  
[61] stringi_1.4.6               XVector_0.26.0              ellipsis_0.3.0              vctrs_0.2.4                 ensembldb_2.10.2           
[66] tools_3.6.3                 bit64_0.9-7                 Biobase_2.46.0              glue_1.4.0                  purrr_0.3.4                
[71] hms_0.5.3                   parallel_3.6.3              AnnotationDbi_1.48.0        GenomicRanges_1.38.0        memoise_1.1.0

The way tximeta works is by reading the hash value that salmon index produces, however you can obtain this without running index. I do this for example when hashing GENCODE and Ensembl. Rob has a python package called fasta-digest you can install with pip that will produce the hash value as a standalone output.