lexical error returned when running makeLinkedTxome

[wzmin]

Hi Mike,

CSHL 2019 participant here! I'm trying to import my quants from Salmon 1.0.0 using whole genome decoys generated on my machine but tximeta couldn't recognize the mouse Gencode M23 that i'm using. I tried to generated my own Txome but got the following error:

makeLinkedTxome(indexDir="/Users/wangz2/Documents/Lab/Computational_resources/salmon_index", source="Gencode", organism="Mus musculus",release="M23", genome="GRCm38", fasta=fastaFTP, gtf=gtfFTP, write=T)

Error: lexical error: invalid char in json text. /Users/wangz2/Documents/Lab/Comp (right here) ------^

This code use to work for me back in August. Any insight into why this is happening?

Thank you, Zhongmin

[mikelove]

Hmm @rob-p i thought that we solved the issue with the decoys breaking the hash? How does the hash get stored in v1.0.0?

The other issue (linkedTxome) is resolved by upgrading jsonlite I believe.

[wzmin]

I just tried upgrading jsonlite and the issue persists. Are we talking about jsonlite 1.6? The same error was reproduced on a different machine.

[mikelove]

Oh also, are you using the current tximeta release, 1.4? Can you post sessionInfo()?

[wzmin]

sessionInfo() R version 3.6.1 (2019-07-05) Platform: x86_64-apple-darwin15.6.0 (64-bit) Running under: macOS Mojave 10.14.6

Matrix products: default BLAS: /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib LAPACK: /Library/Frameworks/R.framework/Versions/3.6/Resources/lib/libRlapack.dylib

locale: [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

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

other attached packages: [1] tximeta_1.2.2

loaded via a namespace (and not attached): [1] Rcpp_1.0.2 lattice_0.20-38 prettyunits_1.0.2 Rsamtools_2.0.3
[5] Biostrings_2.52.0 assertthat_0.2.1 zeallot_0.1.0 digest_0.6.22
[9] BiocFileCache_1.8.0 R6_2.4.0 GenomeInfoDb_1.20.0 backports_1.1.5
[13] stats4_3.6.1 RSQLite_2.1.2 httr_1.4.1 pillar_1.4.2
[17] zlibbioc_1.30.0 rlang_0.4.1 GenomicFeatures_1.36.4 progress_1.2.2
[21] lazyeval_0.2.2 curl_4.2 rstudioapi_0.10 blob_1.2.0
[25] S4Vectors_0.22.1 Matrix_1.2-17 BiocParallel_1.18.1 stringr_1.4.0
[29] ProtGenerics_1.16.0 RCurl_1.95-4.12 bit_1.1-14 biomaRt_2.40.5
[33] DelayedArray_0.10.0 compiler_3.6.1 rtracklayer_1.44.4 pkgconfig_2.0.3
[37] BiocGenerics_0.30.0 tximport_1.12.3 tidyselect_0.2.5 SummarizedExperiment_1.14.1 [41] tibble_2.1.3 GenomeInfoDbData_1.2.1 IRanges_2.18.3 matrixStats_0.55.0
[45] XML_3.98-1.20 crayon_1.3.4 dplyr_0.8.3 dbplyr_1.4.2
[49] GenomicAlignments_1.20.1 bitops_1.0-6 rappdirs_0.3.1 grid_3.6.1
[53] jsonlite_1.6 DBI_1.0.0 AnnotationFilter_1.8.0 magrittr_1.5
[57] stringi_1.4.3 XVector_0.24.0 vctrs_0.2.0 ensembldb_2.8.1
[61] tools_3.6.1 bit64_0.9-7 Biobase_2.44.0 glue_1.3.1
[65] purrr_0.3.3 hms_0.5.2 parallel_3.6.1 yaml_2.2.0
[69] AnnotationDbi_1.46.1 BiocManager_1.30.9 GenomicRanges_1.36.1 memoise_1.1.0

[mikelove]

Can you upgrade to the release first?

Also some jsonlite debugging

https://github.com/jeroen/jsonlite/issues/230

[rob-p]

@mikelove yes, all of the info should be there. @wzmin , can you please post the contents of aux_info/meta_info.json from one of your quant directories?

[wzmin]

@rob-p { "salmon_version": "1.0.0", "samp_type": "none", "opt_type": "vb", "quant_errors": [], "num_libraries": 1, "library_types": [ "IU" ], "frag_dist_length": 1001, "seq_bias_correct": false, "gc_bias_correct": false, "num_bias_bins": 4096, "mapping_type": "mapping", "num_valid_targets": 140748, "num_decoy_targets": 66, "num_eq_classes": 555909, "serialized_eq_classes": false, "eq_class_properties": [ "range_factorized" ], "length_classes": [ 485, 755, 1421, 2773, 101674 ], "index_seq_hash": "6f92253eb7397009ce667653d94538bc1f0bd85fad71e5c45a7395f6cfe07ffe", "index_name_hash": "3de11815d63c55b9242945c8f5ae3500f5b452804c2ffcb5611b83a079ddfb25", "index_seq_hash512": "", "index_name_hash512": "", "index_decoy_seq_hash": "44c0a33f20575470b02707e1ff7f85c8e361b4361b0171a31b130acf1de2c375", "index_decoy_name_hash": "400b996fcf8292decf9f2d7f7f7c2fd08bf6ca188ba0568de00ca4aeb207747b", "num_bootstraps": 0, "num_processed": 33991528, "num_mapped": 30149906, "num_decoy_fragments": 516520, "num_dovetail_fragments": 174621, "num_fragments_filtered_vm": 682108, "num_alignments_below_threshold_for_mapped_fragments_vm": 17732333, "percent_mapped": 88.6982956459033, "call": "quant", "start_time": "Fri Nov 8 21:41:43 2019", "end_time": "Fri Nov 8 21:44:43 2019" }

[wzmin]

@mikelove interestingly when i try to upgrade with both biocmanager::install or install.packages the 1.2.2 release was installed. > BiocManager::install("tximeta") Bioconductor version 3.9 (BiocManager 1.30.9), R 3.6.1 (2019-07-05) Installing package(s) 'tximeta' trying URL 'https://bioconductor.org/packages/3.9/bioc/bin/macosx/el-capitan/contrib/3.6/tximeta_1.2.2.tgz' Content type 'application/x-gzip' length 278913 bytes (272 KB)

downloaded 272 KB

The downloaded binary packages are in /var/folders/0z/k7rhty7s5v32zhtbxlvsh71w3y8z23/T//RtmpFOZQDw/downloaded_packages

[rob-p]

@mikelove looks like the hash entries for 512 are missing, right? That's odd.

[mikelove]

@rob-p my bad everything is fine 6f9... is in the hash table as it should be. False alarm

@wzmin everything should work out of the box but you need to update R first to get latest Bioc (often the case). These are linked. See Bioc install instructions for details.

[wzmin]

@mikelove I had to uninstall everything related to R on my computer to get the latest versions to install properly and now tximeta is recognizing the transcriptome. Sorry for the unnecessary confusion

[rob-p]

@mikelove --- what I mean, is that the 512-bit index seems to be missing from the meta_info.json. That information is present in the index, but looks like it is not properly propagated to meta_info.json. I just pushed a fix in develop. Are we currently using the SHA512 hash, or the SHA256?

[rob-p]

I think we use the 256 right now (so this should not be the cause of the issue), but now the 512 will again be properly propagated in upstream as of the next release --- the omission of this was a regression.

[mikelove]

Got it. Yes we use 256 now. And likely GA4GH will require a new one, but we haven’t settled yet on a spec. Thanks all.