Bioinformatics pipeline and configs for de novo transcriptome assembly of Alzatea verticillata, Cedrela montana, Graffenrieda emarginata, and Handroanthus chrysanthus.
Five tissues —cambium, flowers, flower buds, leaf buds, and leaves— were collected from four Neotropical tree species (Alzatea verticillata, Cedrela montana, Graffenrieda emarginata, Handroanthus chrysanthus). Libraries were sequenced on Illumina NovaSeq 6000 (2×150 bp). Leaf buds and leaves include three biological replicates from the same trees; for other tissues we pooled equal RNA amounts from 2–3 individuals per species. In total, 29 tissue-specific libraries were generated. For detailed information on sampling locations and wet-lab procedures, please refer to the companion article linked to this pipeline (doi). Assembled transcriptomes can be downloaded from Zenodo: doi Analyses were executed on the Xanadú HPC (Institute for Systems Genomics Computational Biology Core - UConn) using SLURM job manager. The command blocks in this README reproduce the exact parameters that were used, following the structure of the runs. The RNAseq_nonmodel tutorial (https://github.com/CBC-UCONN/RNAseq_nonmodel) was followed as the operational reference; for detailed system setup and folder scaffolding, please refer to that tutorial.
Raw RNA-Seq reads are available at the NCBI:
| Species | Tissue | Library | SRA accession | BioSample accession | BioProject |
|---|---|---|---|---|---|
| Handroanthus chrysanthus | Cambium | CAM-POOL | SRR35996128 | SAMN53163668 | PRJNA1359388 |
| Handroanthus chrysanthus | Flower buds | FB-POOL | SRR35996127 | SAMN53163669 | PRJNA1359388 |
| Handroanthus chrysanthus | Leaf buds | H04-B1 | SRR35996126 | SAMN53163670 | PRJNA1359388 |
| Handroanthus chrysanthus | Leaf buds | H06-B1 | SRR35996124 | SAMN53163672 | PRJNA1359388 |
| Handroanthus chrysanthus | Leaf buds | H10-B1 | SRR35996122 | SAMN53163674 | PRJNA1359388 |
| Handroanthus chrysanthus | Leaf | H04-L1 | SRR35996125 | SAMN53163671 | PRJNA1359388 |
| Handroanthus chrysanthus | Leaf | H06-L1 | SRR35996123 | SAMN53163673 | PRJNA1359388 |
| Handroanthus chrysanthus | Leaf | H10-L1 | SRR35996121 | SAMN53163675 | PRJNA1359388 |
| Alzatea verticillata | Cambium | ALZATEA-CAM | SRR36004175 | SAMN53199287 | PRJNA1362998 |
| Alzatea verticillata | Flower buds | ALZATEA-BF | SRR36004174 | SAMN53199288 | PRJNA1362998 |
| Alzatea verticillata | Flowers | ALZATEA-FL | SRR36004173 | SAMN53199289 | PRJNA1362998 |
| Alzatea verticillata | Leaf buds | ALZATEA-2-YH | SRR36004172 | SAMN53199290 | PRJNA1362998 |
| Alzatea verticillata | Leaf buds | ALZATEA-3-YH | SRR36004171 | SAMN53199291 | PRJNA1362998 |
| Alzatea verticillata | Leaf buds | ALZATEA-6-YH | SRR36004170 | SAMN53199292 | PRJNA1362998 |
| Alzatea verticillata | Leaf | ALZATEA-6-LEAF | SRR36004169 | SAMN53199293 | PRJNA1362998 |
| Alzatea verticillata | Leaf | ALZATEA-3-LEAF | SRR36004168 | SAMN53199294 | PRJNA1362998 |
| Alzatea verticillata | Leaf | ALZATEA-2-LEAF | SRR36004167 | SAMN53199295 | PRJNA1362998 |
| Cedrela montana | Cambium | CEDRELA-CAM | SRR36016291 | SAMN53202879 | PRJNA1363130 |
| Cedrela montana | Flower buds | CEDRELA-BF | SRR36016290 | SAMN53202880 | PRJNA1363130 |
| Cedrela montana | Flowers | CEDRELA-FL | SRR36016289 | SAMN53202881 | PRJNA1363130 |
| Cedrela montana | Leaf buds | CEDRELA-3-YH | SRR36016288 | SAMN53202882 | PRJNA1363130 |
| Cedrela montana | Leaf buds | CEDRELA-5-YH | SRR36016287 | SAMN53202883 | PRJNA1363130 |
| Cedrela montana | Leaf buds | CEDRELA-6-YH | SRR36016286 | SAMN53202884 | PRJNA1363130 |
| Cedrela montana | Leaf | CEDRELA-3-L | SRR36016285 | SAMN53202885 | PRJNA1363130 |
| Cedrela montana | Leaf | CEDRELA-6-L | SRR36016284 | SAMN53202886 | PRJNA1363130 |
| Cedrela montana | Leaf | CEDRELA-5-L | SRR36016283 | SAMN53202887 | PRJNA1363130 |
| Graffenrieda emarginata | Flowers | GRAFF-FL | SRR36016315 | SAMN53197954 | PRJNA1362976 |
| Graffenrieda emarginata | Leaf buds | GRAFF-YH | SRR36016314 | SAMN53197964 | PRJNA1362976 |
| Graffenrieda emarginata | Leaf | GRAFF-LEAF | SRR36016313 | SAMN53198155 | PRJNA1362976 |
- Quality Control
- Taxonomic filtering (Kraken2)
- Trimming
- Assembling Transcriptomes
- Identifying the Coding Regions
- Determining and Removing Redundant Transcripts
- Evaluating the Assembly
- Functional Annotation
- Quantifying gene expression
- Final de novo transcriptome datasets
- Reproducibility
- Software references
- Contact
Quality control of Illumina reads with FastQC and aggregation with MultiQC.
From the 02_Quality_Control/ directory:
module load fastqc/0.12.1
# Single sample
SAMPLE="${SPS}_${ID}"
# FastQC → raw_fastqc/
fastqc --outdir ./raw_fastqc/ ../01_Raw_Reads/${SAMPLE}_R1_001.fastq.gz ../01_Raw_Reads/${SAMPLE}_R2_001.fastq.gz
# MultiQC → raw_multiqc/
multiqc --outdir raw_multiqc ./raw_fastqc/Classify reads and split into unclassified (kept for plant assembly) and classified (non-plant) fractions.
Inputs:
${SAMPLE}_R1_001.fastq.gz
${SAMPLE}_R2_001.fastq.gz
Outputs:
${SAMPLE}_unclassified#.fastq
${SAMPLE}_classified#.fastq
${SAMPLE}_report.txt
Script structure for Kraken2:
module load kraken/2.1.2
SAMPLE="${file}"
THREADS=16
KRAKEN2_DB="${KRAKEN2_DB}" # b+a+v+f
mkdir -p kraken_output/kraken_unclassified kraken_output/kraken_classified
kraken2 --db "${KRAKEN2_DB}" \
--paired "${SAMPLE}_R1_001.fastq.gz" "${SAMPLE}_R2_001.fastq.gz" \
--use-names \
--threads "${THREADS}" \
--unclassified-out kraken_output/kraken_unclassified/${SAMPLE}_unclassified#.fastq \
--classified-out kraken_output/kraken_classified/${SAMPLE}_classified#.fastq \
--report kraken_output/${SAMPLE}_report.txtAdapter and quality trimming on Kraken2-unclassified reads with Trimmomatic 0.39.
Inputs:
${SAMPLE}_unclassified_1.fastq
${SAMPLE}_unclassified_2.fastq
Outputs:
Paired: trim_${SAMPLE}_R1.fastq.gz
trim_${SAMPLE}_R2.fastq.gz
Singletons: singles_trim_${SAMPLE}_R1.fastq.gz
singles_trim_${SAMPLE}_R2.fastq.gz
Script structure for Trimmomatic:
module load Trimmomatic/0.39
THREADS=16
ADAPTERS="TruSeq3-PE-2.fa"
java -jar "$Trimmomatic" PE -threads "${THREADS}" \
kraken_output/kraken_unclassified/${SAMPLE}_unclassified_1.fastq \
kraken_output/kraken_unclassified/${SAMPLE}_unclassified_2.fastq \
trim_${SAMPLE}_R1.fastq.gz singles_trim_${SAMPLE}_R1.fastq.gz \
trim_${SAMPLE}_R2.fastq.gz singles_trim_${SAMPLE}_R2.fastq.gz \
ILLUMINACLIP:${ADAPTERS}:2:30:10 \
HEADCROP:10 SLIDINGWINDOW:4:25 MINLEN:45Parameter notes:
ILLUMINACLIP:TruSeq3-PE-2.fa:2:30:10 — removes Illumina adapters using the TruSeq PE list; allows 2 seed mismatches, requires palindrome clip score ≥30 (for overlapping PE reads) and simple clip score ≥10 for standard matches.
HEADCROP:10 — drops the first 10 bases from each read to eliminate low-quality/biased starts.
SLIDINGWINDOW:4:25 — scans with a 4-bp window and trims once the average Phred quality <25 (~0.3% error per base).
MINLEN:45 — discards reads shorter than 45 bp after trimming to avoid spurious mappings/assemblies.
To create a single reference transcriptome for all samples, each sample was separately assembled, resulting transcripts were pooled and clustered into groups that represent single genes. Finally, select a single "best" representative transcript for each gene. De novo transcriptome assembly with Trinity 2.15.0 on trimmed paired reads.
Inputs:
trim_${SAMPLE}_R1.fastq.gz
trim_${SAMPLE}_R2.fastq.gz
(singletons not used here)
Outputs:
trinity_out_dir/Trinity.fasta
Trinity logs and intermediate files in trinity_out_dir/
module load Trinity/2.15.0
SAMPLE="${SAMPLE}"
CPU=16
MEM="100G" # high-mem node
Trinity \
--seqType fq \
--left trim_${SAMPLE}_R1.fastq.gz \
--right trim_${SAMPLE}_R2.fastq.gz \
--CPU ${CPU} \
--max_memory ${MEM} \
--min_contig_length 300 \
--output trinity_${SAMPLE} \
--full_cleanupParameter notes:
--min_contig_length 300 — filters very short contigs to reduce noise before downstream steps.
--max_memory 100G — ensures sufficient RAM for in-memory k-mer graphs on large plant datasets.
We detect long ORFs, annotate PFAM domains on the candidate peptides, and then predict coding regions while retaining PFAM-supported hits.
Inputs:
trinity_combine_${spp}.fasta (assembled transcripts)
Key outputs:
${ASSEMBLY}.transdecoder.gff3
${ASSEMBLY}.transdecoder.bed
Predicted peptides *.transdecoder.pep and CDS *.transdecoder.cds
PFAM domain table: pfam.domtblout
1) Long ORFs scan (TransDecoder.LongOrfs)
module load TransDecoder/5.3.0
SPS="${SPS}"
ASSEMBLY="trinity_combine_${SPS}.fasta"
TransDecoder.LongOrfs -t "${ASSEMBLY}"2) PFAM domain search on candidate peptides (HMMER hmmscan)
module load hmmer/3.2.1
# PFAM HMM library
PFAM_DB="${PFAM_DB}"
CPU="${SLURM_NTASKS:-16}"
hmmscan --cpu "${CPU}" \
--domtblout pfam.domtblout \
"${PFAM_DB}" \
"${ASSEMBLY}.transdecoder_dir/longest_orfs.pep"3) Coding region prediction with PFAM retention (TransDecoder.Predict)
module load TransDecoder/5.3.0
CPU="${SLURM_NTASKS:-16}"
TransDecoder.Predict -t "${ASSEMBLY}" \
--retain_pfam_hits pfam.domtblout \
--cpu "${CPU}"Notes:
TransDecoder.LongOrfs identifies candidate ORFs and writes longest_orfs.pep inside ${ASSEMBLY}.transdecoder_dir/
hmmscan populates pfam.domtblout, which is passed to TransDecoder.Predict via --retain_pfam_hits to prioritize ORFs with conserved domains.
Cluster predicted coding sequences (CDS) to collapse near-duplicate transcripts and retain one representative (“centroid”) per cluster.
Inputs:
trinity_combine_${SPS}.fasta.transdecoder.cds (CDS from TransDecoder)
Outputs:
centroids_${SPS}.fasta (non-redundant CDS / “unigenes”)
clusters.uc (cluster membership table)
LOGFile (vsearch run log)
module load vsearch/2.4.3
SPS="${SPS}"
CPU="${SLURM_NTASKS:-16}"
vsearch \
--threads "${CPU}" \
--log LOGFile \
--cluster_fast trinity_combine_${SPS}.fasta.transdecoder.cds \
--id 0.90 \
--centroids centroids_${SPS}.fasta \
--uc clusters.ucParameter notes:
--cluster_fast — greedy clustering of nucleotide CDS; sequences are compared in their given orientation.
--id 0.90 — merges sequences with ≥90% global identity.
--centroids — writes the representative sequence for each cluster (use as the non-redundant CDS set downstream).
--uc — maps every input sequence to its cluster ID and centroid (traceability).
Assess contiguity/ORF-level metrics with rnaQUAST 1.5.2 (enabling GeneMarkS-T 5.1) and estimate completeness with BUSCO 5.4.5 using the Viridiplantae lineage.
Inputs:
centroids_${SPS}.fasta (non-redundant CDS from clustering)
Outputs:
Genemark_${SPS}/ (rnaQUAST reports, tables, plots; GeneMarkS-T predictions)
busco_out/ (BUSCO summary, full tables, single-copy/duplicated/fragmented/missing counts)
rnaQUAST (with GeneMarkS-T)
module load rnaQUAST/1.5.2
module load GeneMarkS-T/5.1
SPS="${SPS}"
CPU="${SLURM_NTASKS:-16}"
rnaQUAST.py \
--transcripts centroids_${SPS}.fasta \
--gene_mark \
--threads "${CPU}" \
--output_dir Genemark_${SPS}Notes:
--gene_mark enables GeneMarkS-T on transcripts to report ORF-level metrics.
rnaQUAST summarizes N50/L50, lengths, %GC, and ORF statistics in Genemark_${SPS}.
BUSCO (transcriptome mode; Viridiplantae odb10)
module load busco/5.4.5
SPS="${SPS}"
CPU="${SLURM_NTASKS:-16}"
BUSCO_LINEAGE="${BUSCO_LINEAGE:viridiplantae_odb10}"
busco \
-i centroids_${SPS}.fasta \
-o busco_out \
-l "${BUSCO_LINEAGE}" \
-m transcriptome \
-c "${CPU}" \
-fNotes
-m transcriptome expects potentially fragmented gene models.
Annotate the non-redundant peptides with EnTAP, which orchestrates DIAMOND homology searches against curated protein databases, applies taxonomic contaminant filtering, and transfers functional terms (e.g., GO/orthology) from the best-supported hits according to entap_config.ini.
Inputs:
centroids_${SPS}.pep (non-redundant peptide sequences)
entap_config.ini
entap_run.params
DIAMOND databases: RefSeq (complete proteins), Swiss-Prot, and nr (*.dmnd)
Outputs:
EnTAP run directory (e.g., entap_out/) containing summary tables, best-hit assignments, GO/orthology mappings, and logs.
mkdir -p entap_output
module load EnTAP/2.3.0
EnTAP --run \
--entap-ini entap_config.ini \
--run-ini entap_run.params \
--input centroids_${SPS}.pep \
--out-dir entap_output/Parameter notes:
--entap-ini controls coverage/e-value thresholds, taxonomy, and optional expression filters (e.g., FPKM ≥ 0.5).
To further curate the de novo assembly and remove non-plant contaminants, we used the taxonomic classification provided by EnTAP. From the annotated_contam.faa file, which contains proteins flagged as contaminants (fungi, bacteria, archea, insecta), we extracted the corresponding sequence identifiers and used them to filter the centroid transcriptome, removing the associated entries from the centroid FASTA file.
grep -c "^>" 07_EnTAP/entap_output/final_results/annotated_contam.faa
- Extract the IDs of contaminants recorded by EnTAP:
awk '/^>/{gsub(/^>/,""); print $1}' annotated_contam.faa | sort -u > contam_ids.txt
- Clean the transcriptome
seqkit grep -v -f contam_ids.txt centroids_${SPS}.fasta > centroids_${SPS}_filtered.fasta
The resulting clean fasta, depleted of non-plant transcripts, was then used as the reference for expression quantification and all downstream analyses.
Estimate transcript abundances against the plant non-redundant CDS using kallisto 0.46.1. This involves (1) building an index from the clustered CDS FASTA and (2) quantifying each paired library against that index. Singletons were not used.
Input:
centroids_${SPS}_filtered.fasta
Output:
centroids_${SPS}_filtered.fasta.index (kallisto index)
module load kallisto/0.46.1
SPS="${SPS}"
kallisto index -i centroids_${SPS}_filtered.fasta.index centroids_${SPS}_filtered.fastaInputs:
Trimmed reads: trim_${SAMPLE}_R1.fastq.gz
trim_${SAMPLE}_R2.fastq.gz
Index: centroids_${SPS}_filtered.fasta.index
Outputs:
abundance.tsv
abundance.h5
run_info.json
module load kallisto/0.46.1
SPS="${SPS}"
SAMPLE="${SAMPLE}"
CPU="${SLURM_NTASKS:-16}"
kallisto quant \
-i centroids_${SPS}_filtered.fasta.index \
-o "${SAMPLE}" \
-t "${CPU}" \
trim_${SAMPLE}_R1.fastq.gz \
trim_${SAMPLE}_R2.fastq.gzWe removed unigenes with negligible expression, which are likely to represent assembly noise rather than biologically meaningful genes. Unigenes with TPM ≥ 0.5 in at least two samples were retaining for the de novo transcriptome.
| Species | Unigenes | avg_len | max_len |
|---|---|---|---|
| Handroanthus chrysanthus | 114,659 | 687.3 | 15,276 |
| Alzatea verticillata | 98,197 | 647.8 | 14,286 |
| Cedrela montana | 88,586 | 667.9 | 15,546 |
| Graffenrieda emarginata | 91,970 | 605.7 | 12,363 |
-
FastQC 0.12.1; MultiQC 1.9
-
Kraken2 2.1.2 (db: bacteria + archaea + viruses + fungi)
-
Trimmomatic 0.39
-
Trinity 2.15.0
-
TransDecoder 5.3.0; HMMER 3.2.1
-
VSEARCH 2.4.3
-
rnaQUAST 1.5.2 + GeneMarkS-T 5.1
-
BUSCO 5.4.5 (odb10; Viridiplantae lineage)
-
Kallisto v0.46.1
-
EnTAP v2.3.0 with DIAMOND databases:
-
NCBI RefSeq non-redundant protein sequences (Complete genomes, release 232)
-
UniProt UniRef90 protein clusters (release 2025_06)
-
NCBI non-redundant protein database (2025)
-
-
Andrews, S. (2012) FastQC a quality control tool for high throughput sequence data. Available at: http://www.bioinformatics.babraham.ac.uk/projects/fastqc/.
-
Bolger, A.M., Lohse, M. and Usadel, B. (2014) ‘Trimmomatic: A flexible trimmer for Illumina sequence data’, Bioinformatics, 30(15), pp. 2114–2120. Available at: https://doi.org/10.1093/bioinformatics/btu170.
-
Bray, N.L. et al. (2016) ‘Near-optimal probabilistic RNA-seq quantification’, Nature Biotechnology, 34(5), pp. 525–527. Available at: https://doi.org/10.1038/nbt.3519.
-
Bushmanova, E. et al. (2016) ‘rnaQUAST: a quality assessment tool for de novo transcriptome assemblies’, Bioinformatics, 32(14), pp. 2210–2212. Available at: https://doi.org/10.1093/bioinformatics/btw218.
-
Ewels, P. et al. (2016) ‘MultiQC: summarize analysis results for multiple tools and samples in a single report’, Bioinformatics, 32(19), pp. 3047–3048. Available at: https://doi.org/10.1093/bioinformatics/btw354.
-
Grabherr, M.G. et al. (2011) ‘Full-length transcriptome assembly from RNA-Seq data without a reference genome’, Nature Biotechnology, 29(7), pp. 644–652. Available at: https://doi.org/10.1038/nbt.1883.
-
Haas, B.J. et al. (2013) ‘De novo transcript sequence reconstruction from RNA-seq using the Trinity platform for reference generation and analysis’, Nature Protocols, 8(8), pp. 1494–1512. Available at: https://doi.org/10.1038/nprot.2013.084.
-
Hart, A.J. et al. (2020) ‘EnTAP: Bringing faster and smarter functional annotation to non‐model eukaryotic transcriptomes’, Molecular Ecology Resources, 20(2), pp. 591–604. Available at: https://doi.org/10.1111/1755-0998.13106.
-
Mistry, J. et al. (2021) ‘Pfam: The protein families database in 2021’, Nucleic Acids Research, 49(D1), pp. D412–D419. Available at: https://doi.org/10.1093/nar/gkaa913.
-
Rognes, T. et al. (2016) ‘VSEARCH: a versatile open source tool for metagenomics’, PeerJ, 4, p. e2584. Available at: https://doi.org/10.7717/peerj.2584.
-
Simão, F.A. et al. (2015) ‘BUSCO: assessing genome assembly and annotation completeness with single-copy orthologs’, Bioinformatics, 31(19), pp. 3210–3212. Available at: https://doi.org/10.1093/bioinformatics/btv351.
-
Wood, D.E., Lu, J. and Langmead, B. (2019) ‘Improved metagenomic analysis with Kraken 2’, Genome Biology, 20(1), p. 257. Available at: https://doi.org/10.1186/s13059-019-1891-0.
Cristina Zamora Ballesteros (cristinazamoraballesteros@gmail.com).