Simultaneous targeted amplicon deep sequencing and library preparation for a time and cost-effective universal parasite diagnostic sequencing approach

Abstract

We recently described a targeted amplicon deep sequencing (TADS) strategy that utilizes a nested PCR targeting the 18S rDNA gene of blood-borne parasites. The assay facilitates selective digestion of host DNA by targeting enzyme restriction sites present in vertebrates but absent in parasites. This enriching of parasite-derived amplicon drastically reduces the proportion of host-derived reads during sequencing and results in the sensitive detection of several clinically important blood parasites including Plasmodium spp., Babesia spp., kinetoplastids, and filarial nematodes. Despite these promising results, high costs and the laborious nature of metagenomics sequencing are prohibitive to the routine use of this assay in most laboratories. We describe and evaluate a new metagenomic approach that utilizes a set of primers modified from our original assay that incorporates Illumina barcodes and adapters during the PCR steps. This modification makes amplicons immediately compatible with sequencing on the Illumina MiSeq platform, removing the need for a separate library preparation, which is expensive and time-consuming. We compared this modified assay to our previous nested TADS assay in terms of preparation speed, limit of detection (LOD), and cost. Our modifications reduced assay turnaround times from 7 to 5 days. The cost decreased from approximately $40 per sample to $11 per sample. The modified assay displayed comparable performance in the detection and differentiation of human-infecting Plasmodium spp., Babesia spp., kinetoplastids, and filarial nematodes in clinical samples. The LOD of this modified approach was determined for malaria parasites and remained similar to that previously reported for our earlier assay (0.58 Plasmodium falciparum parasites/μL of blood). These modifications markedly reduced costs and turnaround times, making the assay more amenable to routine diagnostic applications.

Background

We recently described a universal parasite diagnostic assay (UPDx), a metagenomic assay targeting the eukaryotic 18S rDNA gene that was developed to characterize parasitic communities in blood samples (Flaherty et al. 2018). This first UPDx assay was later modified to include a nested amplification step that improved the assay’s limit of detection (LOD) (Flaherty et al. 2021). This nested version of UPDx was referred to as nUPDx (Flaherty et al. 2021). Unique among metagenomics assays, the nUPDx amplicon possesses restriction sites that exist only in vertebrates and not in parasites. Taking advantage of these vertebrate-specific restriction sites, nUPDx includes restriction enzyme digestion steps that are performed on the genomic DNA extract prior to amplification and on the PCR product from the first amplification. These digestion steps reduce the abundance of host-derived sequences in the final PCR product and therefore result in markedly improved sensitivity for the identification of any parasites in the samples (Flaherty et al. 2021).

The nUPDx assay has a limit of detection (LOD) similar to pathogen-specific real-time PCR assays and facilitated detection of DNA from malaria parasites (Plasmodium malariae, P. falciparum, P. vivax, and P. ovale), Babesia species (B. microti, B. divergens, and B. duncani), kinetoplastids (Leishmania spp., Trypanosoma cruzi, and Trypanosoma brucei), and the filarial nematodes Loa loa and Brugia malayi (Flaherty et al. 2021). In most cases, it was able to provide a species-level diagnosis, which can be critical for proper patient management, such as determining the appropriate drug treatment for malaria. Despite its utility, as with other metagenomic sequencing methods, the library preparation protocol required for nUPDx is laborious, cumbersome, and very expensive in terms of time (person hours) and monetary costs (Hess et al. 2020). These aspects represent a significant barrier to the routine diagnostic use of nUPDx and similar metagenomic assays in place of cheaper and widely available PCR-based diagnostics.

To counteract this barrier, Illumina published a guide describing how investigators can combine PCR amplification with library preparation to produce sequencingready amplicons, thereby reducing the need for library preparation (Anonymous 2022). Briefly, these guidelines describe how to modify user‐defined PCR primers, allowing amplification of target DNA while simultaneously preparing the amplicon for sequencing by incorporating the necessary Illumina adapter sequences. A subsequent amplification step is then performed with primers complementary to adapter sequences, to add multiplexing indices and sequencing adapters for MiSeq system sequencing. This protocol and its adaptations have already been successfully applied to metagenomics studies (Lee et al. 2019; Diaz-Torres et al. 2021), including an adaptation of nUPDx developed at the Wadsworth Center in the New York State Department of Health (Clemons et al. 2022).

Given the diagnostic utility displayed by nUPDx, we sought to build upon the work of the Wadsworth Center (Clemons et al. 2022) by improving nUPDx further while also removing the need for expensive library preparation. In the Wadsworth Center study, as samples still underwent a nested PCR with the addition of overhang adapters during the second amplification step, a third amplification step was required to add indices and sequencing adapters from Illumina. In this study, we shortened the procedure to only two amplification steps by adding overhang adapters in the first step and custom index-incorporating primers in the second step. This study sought to compare this improved adapter-incorporating UPDx method (Ad_UPDx) to the previously described nUPDx approach (Flaherty et al. 2021) in terms of preparation speed, LOD, and cost.

We applied Ad_UPDx to various clinical blood samples containing a range of blood parasites including the apicomplexan parasites Plasmodium spp. and Babesia spp., which are nationally notifiable in the USA (Hwang et al. 2009; Bishop et al. 2021). Parasites from these genera cause potentially lethal infections, and the incidence of malaria and babesiosis diagnoses is increasing in the USA due to increasing international travel and an increase in domestically acquired infections, respectively (Dye-Braumuller and Kanyangarara 2021; Mace et al. 2021; Menis et al. 2021). These trends highlight the need for modern diagnostic assays that accurately detect and differentiate morphologically similar bloodborne parasites.

Additionally, we tested blood samples containing kinet-oplastid parasites and filarial nematodes. These parasites are encountered less frequently in the USA than Plasmodium spp. and Babesia spp., although detection and differentiation of rarer infections are also important functions of reference diagnostic laboratories where we anticipate that Ad_UPDx could be implemented. We also assessed the LOD of Ad_UPDx using serially diluted, quantified cultures of P. falciparum.

Methods

Source of samples

A total of 36 blood samples were analyzed in parallel using nUPDX and Ad_UPDx for direct comparison. Blood samples were confirmed positive or negative for parasites using pathogen-specific PCR and/or by light microscopic examination of stained blood smears, as described in Table 1. Thirteen samples had tested positive for a single parasite, which included Plasmodium falciparum, Plasmodium ovale, Plasmodium malariae, Plasmodium vivax, Babesia microti, Babesia divergens-like variant MO1, Leishmania sp., Brugia malayi, Loa Loa (n = 1 each), and Trypanosoma cruzi collected during acute infection (n = 2) and during chronic infection (n = 2). Three samples had tested positive for multiple malaria parasites, including P. falciparum/P. ovale, P. falciparum/P. vivax, and P. falciparum/P. malariae. To determine the LOD of Ad_UPDx, we tested a serial dilution of a Plasmodium falciparum strain 3D7 culture spiked into parasite-free blood in duplicate with concentrations ranging from 58,000 to 0.0058 parasite/μL (i.e., eight tenfold dilution steps). P. falciparum was chosen as the representative parasite for LOD estimation for Ad_UPDx because it had been used for this purpose for nUPDx previously (Flaherty et al. 2021). Parasite culture and serial dilutions were prepared as before (Flaherty et al. 2021) to produce two sets of serial dilutions (duplicates). For this study, the LOD was defined as the lowest concentration that generated positive results in at least one of the duplicates. Finally, four parasite-free samples from healthy blood donors were included in this study as negative controls.

Table 1.

Description of blood samples used in this study including source, identification methods, and host

Parasite	Sample type	Host	Species identification diagnostic method/s [reference]	Source
Plasmodium falciparum	EDTA blood	Homo sapiens	Microscopy and real-time PCR (Rougemont et al. 2004)	CDC
Plasmodium falciparum strain 3d7	Culture	Homo sapiens	Microscopy	CDC
Plasmodium vivax	EDTA blood	Homo sapiens	Microscopy and real-time PCR (Rougemont et al. 2004)	CDC
Plasmodium malariae	EDTA blood	Homo sapiens	Microscopy and real-time PCR (Rougemont et al. 2004)	CDC
Plasmodium ovale	EDTA blood	Homo sapiens	Microscopy and real-time PCR (Rougemont et al. 2004)	CDC
Mixed infection P. falciparum and P. ovale	EDTA blood	Homo sapiens	Microscopy and PCR (Snounou et al. 1993)	CDC
Mixed infection P. falciparum and P. malariae	EDTA blood	Homo sapiens	Microscopy and PCR (Snounou et al. 1993)	CDC
Mixed infection P. falciparum and P. vivax	EDTA blood	Homo sapiens	Microscopy and PCR (Snounou et al. 1993)	CDC
Trypanosoma cruzi (from chronic infection – samples R1 and R2)	EDTA blood	Homo sapiens	Real-time PCR (Qvarnstrom et al. 2012)	CDC
Trypanosoma cruzi (from acute infection – samples A1 and A2)	EDTA blood	Homo sapiens	Real-time PCR (Qvarnstrom et al. 2012)	Institute of Public Health, Chile
Leishmania sp.	EDTA blood	Homo sapiens	Microscopy and PCR (Almeida et al. 2021)	CDC
Babesia divergens-like variant MO1	EDTA blood	Homo sapiens	Microscopy and PCR (Bonnet et al. 2007)	CDC
Babesia microti	EDTA blood	Homo sapiens	Microscopy and real-time PCR (Souza et al. 2016)	CDC
Loa Loa	EDTA blood	Homo sapiens	Microscopy	CDC
Brugia malayi	Whole blood	Felis catus	Microscopy	University of Georgia, USA
NPF (no parasite found)	EDTA blood	Homo sapiens	Microscopy and PCR (Qvarnstrom et al. 2012; Rougemont et al. 2004; Almeida et al. 2021; Bonnet et al. 2007; Souza et al. 2016)	CDC

EDTA, ethylenediaminetetraacetic acid; CDC, Centers for Disease Control and Prevention

DNA extraction

DNA was extracted using a QIAamp DNA Blood Mini QIACube Kit on a QIACube for automated extraction, according to the manufacturer’s instructions (QIAGEN, Germantown, MD, USA). The elution volume of Buffer AE was adjusted to 50 μL.

Ad_UPDx assay design

The Ad_UPDx assay was designed by combining the previously described nUPDx assay (Flaherty et al. 2021) and the NEBNext Ultra DNA Library Prep Kit protocol for Illumina sequencing (New England Biolabs, Ipswich, MA, USA). The first amplification targeted a ~ 200 bp fragment of the 18S rRNA genes using primers that possessed overhang adapters containing the priming sites for a second amplification (Fig. 1). The second amplification was then performed with primers complementary to the adapter overhang sequences incorporated during the first PCR, allowing nested amplification and addition of indices and flow cell adapters simultaneously (Fig. 1).

Fig. 1.

Combined DNA amplification and library preparation using the newly described Ad_UPDx protocol. First, specific and user-defined forward and reverse primers targeting the region of interest are designed with overhang adapters compatible with Illumina sequencing and used for template amplification from DNA samples. Then, primers complementary to adapter sequences are used to attach multiplexing indices and Illumina flow cell adapters

Forward and reverse UPDx primers with overhang adapters

Overhang adapter sequences were added to the 5’ end of the inner nUPDx primers. Adapter sequences were modified from the NEBNext^® Multiplex Oligos for Illumina^® (96 Index Primers) manual for Illumina sequencing (New England Biolabs). Sequences of the newly designed forward and reverse Ad_UPDx primers are available in Table 2.

Table 2.

Ad_UPDx Primer sequences designed in this study

Primer name	Primer sequences (5’−3’)
AdUPDx_F1 (UPDx inner forward + Adapter1)	GTG ACT GGA GTT CAG ACG TGT GCT CTT CCG ATC TCC GGA GAG GGAG CC TGAGA
AdUPDx_R1 (UPDx inner reverse + Adapter2)	ACA CTC TTT CCC TAC ACG ACG CTC TTC CGA TCT GAG CTG GAA TTA CCG CGG
Index_PrimerF (Sequence complementary to Oligoflowcell1 + INDEX + Sequence complementary to Adapter2)	CAA GCA GAA GAC GGC ATA CGA GAT *GTG ACT GGA GTT CAG ACG TGT GCT CTT CCG ATCT
Univ_PrimerR (Sequence complementary to Oligoflowcell2 + Sequence complementary to Adapter1)	AAT GAT ACG GCG ACC ACC GAG ATC TAC ACT CTT TCC CTA CAC GAC GCT CTT CCG ATCT

The sequences were adapted from Flaherty et al. (2021) and NEBNext^® Multiplex Oligos for Illumina^® 96 Index Primers E6609 manual (available at https://international.neb.com//media/nebus/files/manuals/manuale6609.pdf)

Preparation of primers for index and universal adapter-incorporating PCR

Primer sequences for the second amplification step were adapted from the NEBNext^® Multiplex Oligos for Illumina^® (96 Index Primers) manual for Illumina sequencing (New England Biolabs). A total of 96 forward index-incorporating primers were designed along with one reverse primer (Table 2, Supplementary information). To simplify reaction preparation, primer mixes were prepared in 96-well microplates. Each well contained 7 μL of one of the 10 μM forward index primers and 7 μL of 10 μM reverse primer (Univ_primerR).

First enzymatic digestion

Extracted DNA (7.5 μL) was subjected to the first enzymatic digestion in a final volume of 10 μL, including 0.5 μL of BamHI-HF (10 units), 1 μL of BsoBI (10 units), and 1 μL of 10X CutSmart Buffer. Samples were incubated for 1 h at 37 °C. All restriction enzymes were purchased from New England Biolabs.

First amplification step using UPDx primers with overhang adapters

Digested DNA (2 μL) was amplified using the AdUPDx_F1 and AdUPDx_R1 primers. Each reaction included 5 μL of 5X Q5 Reaction Buffer, 5 μL of 5X High GC Enhancer, 0.25 μL (500 units) of Q5 High-Fidelity DNA Polymerase, 0.5 μL of dNTPs (solution mix of 10 mM each), 9.5 μL PCR grade water, and 1.25 μL of each 10 μM primer, in a final volume of 25 μL. All PCR reagents (excluding primers and water) were purchased from New England Biolabs. Thermal cycling was performed as follows: 98.0 °C for 30 s, 30 cycles of 98.0 °C for 10 s, 67.0 °C for 30 s, 72.0 °C for 45 s, and 72.0 °C for 2 min.

Second enzymatic digestion

Restriction enzymes and buffer were directly added to the 25 μL of PCR products as follows: 0.5 μL of BamHI-HF (10 units), 1 μL of BsoBI (10 units), and 2.5 μL of 10X CutSmart Buffer. Samples were again incubated for 1 h at 37 °C.

Bead cleanup with size selection

The digested PCR product (29 μL) was processed according to NEBNext^® Ultra^™ II DNA Library Prep Kit manufacturer instructions (New England Biolabs) by performing a bead cleanup with size selection using Agencourt AMPure XP Beads (Beckman Coulter, Brea, CA, USA), accounting for an approximate insert size of 200 bp. Amplicons were eluted in 15 μL of 0.1X TE buffer.

Second amplification step for the addition of multiplexing indices and sequencing adapters

Cleaned amplicons (15 μL) were subjected to the second amplification step using the forward index and reverse universal primers (Table 2, Supplementary information). Reactions were prepared to contain 25 μL of NEBNext Ultra II Q5 Master Mix (New England Biolabs) and 10 μL of the primer mix previously prepared in the 96 microwell plates, including one forward index primer and the reverse Universal primer. Next, 15 μL of cleaned amplicon was added to the mix to a final volume of 50 μL. Thermal cycling was performed as follows: 98.0 °C for 30 s, 10 cycles of 98.0 °C for 10 s, 65.0 °C for 75 s, and 65.0 °C for 5 min.

nUPDx assay

The same DNA samples were processed with the comparison method, nUPDx, (Flaherty et al. 2021) in parallel to the Ad_UPDx method. During the library preparation index PCR step, index primers were selected that had not already been used with the Ad_UPDx method so that the same samples detected with both methods could be multiplexed in the same sequencing run.

Bead cleanup without size selection

A final bead cleanup without size selection was performed on sample libraries according to NEBNext^® Ultra^™ II DNA Library Prep Kit manufacturer instructions (New England Biolabs) using Agencourt AMPure XP Beads (Beckman Coulter). The amplicons were eluted in 40 μL of 0.1X TE buffer, and 5 μL of eluate from each sample was pooled for each method (Ad_UPDx and nUPDx) separately, resulting in two pooled libraries. Library fragment size estimation and DNA concentration were determined as described below.

Library fragment size and concentration

The concentration of the final pooled libraries for nUPDx and Ad_UPDx were individually determined using a Qubit 2.0 Fluorometer with the Qubit dsDNA High Sensitivity Assay Kit (Life Technologies, Grand Island, NY, USA). The average fragment size of each sample library pool was determined by the Agilent 2200 Tapestation System using Agilent D1000 ScreenTape reagents following manufacturer instructions (Agilent Technologies, Santa Clara, CA, USA).

Amplicon sequencing

Pooled libraries constructed using the Ad_UPDx and nUPDx methods were separately denatured and diluted to 10 pM. This normalization was performed to ensure that reads from each pooled library would be represented equally in the sequencing run. The pooled library was then spiked with the 10% PhiX control library (Illumina, San Diego, CA, USA), as per the manufacturer’s recommendations. The pooled libraries were sequenced using MiSeq^® Reagent Nano Kit v2 (500 cycles) (Illumina) on an Illumina MiSeq Sequencing platform following the manufacturer’s instructions.

Bioinformatic analysis

All raw sequencing reads have been made publicly available on the NCBI Sequence Read Archive under BioProject accession number PRJNA437674. Sequencing data were analyzed using a custom bioinformatic workflow. This workflow was adapted from open-source resources (Callahan et al. 2016; Lee 2019) and is available at the following GitHub repository: https://github.com/Mathilg/UPDx_workflow.git. Briefly, after primer removal using Cutadapt V2.10, reads were processed using the Dada2 R package (R version 4.0) and RStudio (version 1.4.1106). Reads were filtered based on quality, where sequences were truncated at the first instance of a quality score less than 2, and after truncating, sequences with an overall quality score less than 15 or with more than 2 potential erroneous base calls were discarded. Trimmed reads were filtered by length, where only reads ranging from 145 to 250 bases long were retained. Reads were then de-replicated (with a parameter of 100% sequence identity), and Amplicon Sequence Variants (ASVs) were determined using the core sample inference algorithm of Dada2 (Callahan et al. 2016). ASVs were then merged, with a minimum overlap region of 150 bp and 100% sequence identity. Finally, likely chimeras were identified and removed using the Dada2 chimera removal feature. Taxonomic assignment of the final sequences was performed by nucleotide similarity search using BLASTN.

To determine a “positivity” threshold (the number of parasite-matching reads required to validate the presence of that parasite in the sample), we utilized the same cutoff system as for the original UPDx method (Flaherty et al. 2018). Briefly, for a sample to be considered positive for any parasite taxon, the number of reads matching the 18S rDNA of this taxon must exceed either a minimum threshold of 20 reads or a dynamic threshold — whichever value was largest. The dynamic cutoff was computed based on the proportion of parasite-derived reads detected in the negative control samples sequenced in the same library due to index crosstalk, which is an artifact introduced during sample multiplexing (Flaherty et al. 2018). Computing the percentage of parasite-matching reads in a given sample was conducted in R (version 4.0) and RStudio (version 1.4.1106), where the number of reads matching any given parasite sequence was divided by the total number of reads obtained for that sample after cleaning and merging of the sequence data, multiplied by 100.

Results

Detection of various parasite taxa in blood

The diversity of parasite species detected using Ad_UPDx and nUPDx was the same based on the present comparison (Table 3, Fig. 2). The taxonomic assignment of sequences identified in blood samples and the proportion of parasite-derived reads obtained for each method is shown in Table 3 and Fig. 2.

Table 3.

Percentage of parasite-matching reads and taxonomic assignment using BLASTN searches of the sequences detected by both methods (Ad_UPDx and nUPDx)

Sample	Sequence detected	Parasite vs. total reads by Ad_UPDx (%)	Parasite vs. total reads by nUPDx (%)	Sequence length (bp)	Closest homology	Id%	GB number
Trypanosoma cruzi (A1)	ASV_Trypano-soma_cruzi	996/1284 (77.6%)	751/753 (99.8%)	179	Trypanosoma cruzi	100	KX007998.1
Trypanosoma cruzi (A2)	ASV_Trypano-soma_cruzi	1372/1476 (93%)	881/887 (99.4%)	179	Trypanosoma cruzi	100	KX007998.1
Trypanosoma cruzi (R1)	ASV_Trypano-soma_cruzi	898/1514 (59.4%)	460/1146 (40.2%)	179	Trypanosoma cruzi	100	KX007998.1
Trypanosoma cruzi (R2)	No parasite-derived sequence detected	0/2222	0/2919	N/A	N/A	N/A	N/A
Babesia microti	ASV_Babesia_microti	5452/5661 (96.4%)	2628/2631 (99.9%)	169	Babesia microti	100	KY649348.1
Babesia divergens-like sp. MO1	ASV_Babesia_spp.	2950/4015 (73.5%)	2697/2776 (97.2%)	168	Babesia spp. (including B. odocoilei, B. capreoli, B. divergens, Babesia venatorum, Babesia sp. MO1)	100	KY805843.1, KY805834.1, MG344781.1, MG344777.1, AY048113.1
Loa loa	ASV_Onchocerci-dae_spp.	1509/2412 (62.6%)	1003/3854 (26.1%)	171	Onchocercidae spp. (including Loa loa)	100	LC378874.1, XR_002251421.1
Brugia malayi	ASV_Filarioidea_spp.	1732/2579 (67.2%)	351/2241 (15.7%)	171	Filarioidea spp.	100	MK868471.1, AF036588.1
Leishmania sp.	ASV_Trypanoso-matidae_spp.	484/1593 (30.4%)	187/2151 (8.7%)	178	Trypanosomatidae spp.	100	MN215474.1, CP040155.1
Plasmodium malariae	ASV_Plasmo-dium_malariae	4740/5519 (85.9%)	1395/2101 (66.4%)	172	Plasmodium malariae	100	KU510228.1
Plasmodium falciparum	ASV_Plasmo-dium_falciparum_1	3737/6227 (60.1%)	533/3390 (15.8%)	173	Plasmodium falciparum	100	LR131493.1
	ASV_Plasmo-dium_falciparum_2	2307/6227 (37.1%)	2770/3390 (81.8%)	173	Plasmodium falciparum	100	MG725888.1
Plasmodium ovale	ASV_Plasmo-dium_ovale_1	1269/2725 (46.6%)	612/2317 (26.5%)	172	Plasmodium ovale	100	MG847138.1
	ASV_Plasmo-dium_ovale_2	379/2725 (14%)	0/2317	173	Plasmodium ovale	100	KY073344.1
Plasmodium vivax	ASV_Plasmo-dium_vivax_1	1271/5489 (23.2%)	2349/3908 (60.2%)	170	Plasmodium spp. (including P. vivax)	100	MK078096.1, XR_003001206.1
	ASV_Plasmo-dium_vivax_2	3223/5489 (58.8%)	864/3908 (22.2%)	170	Plasmodium vivax	100	MT710336.1
Mixed infection with P. falciparum and P. malariae	ASV_Plasmo-dium_falciparum_1	3737/6227 (60.1%)	0/3811	173	Plasmodium falciparum	100	LR131493.1
	ASV_Plasmo-dium_falciparum_2	2307/6227 (37.1%)	352/3811 (9.3%)	173	Plasmodium falciparum	100	MG725888.1
Mixed infection with P. falciparum and P. ovale	ASV_Plasmodium_ovale_1	3415/5476 (62.4%)	2806/3227 (87%)	172	Plasmodium ovale	100	MG847138.1
	ASV_Plasmodium_ovale_2	1321/5476 (24.2%)	65/3227 (2.1%)	173	Plasmodium ovale	100	KY073344.1
Mixed infection with P. falciparum and P. vivax	ASV_Plasmo-dium_falciparum_1	3737/6227 (60.1%)	166/3071 (5.5%)	173	Plasmodium falciparum	100	LR131493.1
	ASV_Plasmo-dium_falciparum_2	2307/6227 (37.1%)	1448/3071 (47.2%)	173	Plasmodium falciparum	100	MG725888.1

For each sample, the number of reads matching the expected target and the total number of reads are listed. The percentage of parasite-matching reads was computed as the number of reads matching any given parasite sequence divided by the total number of reads obtained for that sample after cleaning and merging the sequence data, multiplied by 100. A taxonomic assignment was obtained using BLASTN searches. Id: Percentage of sequence identity when compared to a sequence (or sequences) with the closest homology

Fig. 2.

Percentage of reads matching the target among parasite-positive samples processed with Ad_UPDx and nUPDx methods. After taxonomic assignments of the reads, for each sample infected with at least one parasite, we determined the percentage of reads matching the expected target using nUPDx (blue) or Ad_UPDx (red) methods. This percentage was computed as the number of reads matching any given parasite sequence divided by the total number of reads obtained for that sample after cleaning and merging the sequence data, multiplied by 100. Only one target has been detected in samples containing mixed infections of malaria species and is represented with a star (*)

Reads identical to a T. cruzi reference sequence from GenBank (KX007998.1) (Table 3) were detected in three of the four T. cruzi-positive samples using both nUPDx and Ad_UPDx. Similarly, reads identical to an 18S sequence from Babesia microti (KY649348.1) were detected in the B. microti-positive sample via both methods. For the B. divergens-positive blood sample, we detected reads possessing 100% identity to several Babesia species at the amplified region, including B. odocoilei, B. capreoli, B. divergens, B. venatorum, and Babesia sp. MO1 (KY805843.1, KY805834.1, MG344781.1, MG344777.1, AY048113.1) (Table 3). Reads possessing 100% identity to the expected 18S target for several members of the Onchocercidae family of nematodes and the Filarioidea superfamily of nematodes were detected in samples positive for Loa loa and Brugia malayi, respectively (Table 3), via both methods. For the Leishmania-positive blood sample, reads consistent with the presence of parasites from the Leishmaniinae subfamily were detected using both methods (Table 3).

Plasmodium malariae was detected using both methods, where reads identical to the expected amplicon for reference sequence KU510228.1 were observed. For blood samples containing either P. falciparum, P. vivax, or P. ovale, two different sequences were detected using Ad_UPDx. These sequences correspond to multiple paralogs of the 18S rDNA encoded in the genomes of these Plasmodium species (Steenkeste et al. 2009; Gruenberg et al. 2018). The nUPDx assay produced similar results, with the exception of the P. ovale sample, where nUPDx detected only one of the two 18S rDNA paralogs (Table 3).

For samples comprising naturally acquired mixed malaria species infections, only one of the two expected parasites was detected via both methods in all instances (Table 3). Only sequences corresponding to Plasmodium falciparum were detected in the Plasmodium falciparum/P. malariae and P. falciparum/P. vivax mixed samples. Only sequences corresponding to Plasmodium ovale were detected in the Plasmodium falciparum/P. ovale mixed sample (Table 3, Fig. 2).

One sample, Trypanosoma cruzi R2 (a sample collected from a chronic-phase Chagas disease patient and tested positive by real-time PCR), tested negative for parasites via both methods.

Assessing the limit of detection via serially diluted P. falciparum culture

We analyzed parasite-free human blood spiked with serial dilutions of cultured P. falciparum 3D7 parasites. No parasite-derived sequences were detected at dilutions below 0.58 parasites/μL for both methods (Table 4, Fig. 3). Reads corresponding to the two 18S paralogs expected for P. falciparum were detected in most duplicates, except for duplicate dilutions of 0.58 parasites/μL tested via nUPDx, where only one paralog was detected for one duplicate and the other duplicate was negative for any parasite-derived sequences. Both duplicates of the 0.58 parasites/μL dilution tested via Ad_UPDx were positive, and both expected P. falciparum 18S paralogs were detected in each (Table 4, Fig. 3). Thus, although both methods displayed an LOD of 0.58 parasites/μL; results were more consistent for Ad_UPDx at lower dilutions.

Table 4.

Sequences abundance and taxonomy among serial dilutions of P. falciparum 3D7 parasites in whole human blood

Sample	Sequence name	Ad_UPDx abundance (%)	nUPDx abundance (%)	Sequence length (bp)	Closest homology	Id%	GB number
58,000 parasites/μL (replicate_1)	ASV_Plasmodium_falciparum_1	3424/6060 (56.6%)	540/3459 (15.7%)	173	Plasmodium falciparum	100	LR131493.1
	ASV_Plasmodium_falciparum_2	2217/6060 (36.6%)	2625/3459 (75.9%)	173	Plasmodium falciparum	100	MG725888.1
58,000 parasites/μL (replicate_2)	ASV_Plasmodium_falciparum_1	6406/10734 (59.7%)	699/4570 (15.3%)	173	Plasmodium falciparum	100	LR131493.1
	ASV_Plasmodium_falciparum_2	3717/10734 (34.7%)	3455/4570 (75.7%)	173	Plasmodium falciparum	100	MG725888.1
5800 parasites/μL (replicate_1)	ASV_Plasmodium_falciparum_1	No sequencing results	578/3742 (15.5%)	173	Plasmodium falciparum	100	LR131493.1
	ASV_Plasmodium_falciparum_2	No sequencing results	2879/3742 (77%)	173	Plasmodium falciparum	100	MG725888.1
5800 parasites/μL (replicate_2)	ASV_Plasmodium_falciparum_1	3325/5605 (59.4%)	486/3139 (15.5%)	173	Plasmodium falciparum	100	LR131493.1
	ASV_Plasmodium_falciparum_2	1931/5605 (34.5%)	2435/3139 (77.6%)	173	Plasmodium falciparum	100	MG725888.1
580 parasites/μL (replicate_1)	ASV_Plasmodium_falciparum_1	2009/3706 (54.3%)	473/4449 (10.7%)	173	Plasmodium falciparum	100	LR131493.1
	ASV_Plasmodium_falciparum_2	1249/3706 (33.8%)	2850/4449 (64.1%)	173	Plasmodium falciparum	100	MG725888.1
580 parasites/μL (replicate_2)	ASV_Plasmodium_falciparum_1	2205/4158 (53.1%)	419/3925 (10.7%)	173	Plasmodium falciparum	100	LR131493.1
	ASV_Plasmodium_falciparum_2	1363/4158 (32.8%)	2493/3925 (63.6%)	173	Plasmodium falciparum	100	MG725888.1
58 parasites/μL (replicate_1)	ASV_Plasmodium_falciparum_1	1968/4253 (46.3%)	152/4603 (3.4%)	173	Plasmodium falciparum	100	LR131493.1
	ASV_Plasmodium_falciparum_2	1210/4253 (28.5%)	1241/4603 (27%)	173	Plasmodium falciparum	100	MG725888.1
58 parasites/μL (replicate_2)	ASV_Plasmodium_falciparum_1	1834/3716 (49.4%)	179/4913 (3.7%)	173	Plasmodium falciparum	100	LR131493.1
	ASV_Plasmodium_falciparum_2	1126/3716 (30.4%)	1960/4913 (39.9%)	173	Plasmodium falciparum	100	MG725888.1
5.8 parasites/μL (replicate_1)	ASV_Plasmodium_falciparum_1	1045/3587 (29.2%)	0/493 (0%)	173	Plasmodium falciparum	100	LR131493.1
	ASV_Plasmodium_falciparum_2	689/3587 (19.3%)	196/2265 (8.7%)	173	Plasmodium falciparum	100	MG725888.1
5.8 parasites/μL (replicate_2)	ASV_Plasmodium_falciparum_1	807/2488 (32.5%)	0/1940 (0%)	173	Plasmodium falciparum	100	LR131493.1
	ASV_Plasmodium_falciparum_2	552/2488 (22.2%)	368/1940 (19%)	173	Plasmodium falciparum	100	MG725888.1
0.58 parasites/μL (replicate_1)	ASV_Plasmodium_falciparum_1	157/2045 (7.7%)	0/1047 (0%)	173	Plasmodium falciparum	100	LR131493.1
	ASV_Plasmodium_falciparum_2	155/2045 (7.6%)	57/1047 (5.5%)	173	Plasmodium falciparum	100	MG725888.1
0.58 parasites/μL (replicate_2)	ASV_Plasmodium_falciparum_1	263/2657 (9.9%)	0/2458 (0%)	173	Plasmodium falciparum	100	LR131493.1
	ASV_Plasmodium_falciparum_2	101/2657 (3.9%)	0/2458 (0%)	173	Plasmodium falciparum	100	MG725888.1
0.058 parasites/μL (replicate_1)	No parasite-derived sequence detected	0/2254 (0%)	0/2725 (0%)	N/A	N/A	N/A	N/A
0.058 parasites/μL (replicate_2)	No parasite-derived sequence detected	0/2562(0%)	0/1537(0%)	N/A	N/A	N/A	N/A
0.0058 parasites/μL (replicate_1)	No parasite-derived sequence detected	0/2298 (0%)	0/2570 (0%)	N/A	N/A	N/A	N/A
0.0058 parasites/μL (replicate_2)	No parasite-derived sequence detected	0/2362(0%)	0/2031 (0%)	N/A	N/A	N/A	N/A

For each sample, the number of reads matching the expected target and the total number of reads are listed. The percentage of parasite-matching reads was computed as the number of reads matching any given parasite 18S sequence divided by the total number of reads obtained for that sample after cleaning and merging the sequence data, multiplied by 100. A taxonomic assignment was obtained using BLASTN searches. Id: Percentage of sequence identity when compared to a sequence (or sequences) with the closest homology. GB number: accession number of the reference on Genbank

Fig. 3.

Mean percentage of total reads matching the target among serial dilutions of Plasmodium falciparum 3D7 culture, processed in duplicates with both Ad_UPDx and nUPDx. After cleaning, merging and taxonomic assignments of the reads, the mean percentage of reads matching the expected target (Plasmodium falciparum) was determined for nUPDx (blue) or Ad_UPDx (red) methods by dividing the matching reads by the total number of reads obtained for that sample, multiplied by 100. Mean and standard deviation were computed for each duplicated sample and displayed as error bars

One of two replicates of the second-highest P. falciparum dilution (5800 parasites/μL) analyzed using Ad_UPDx failed to produce a parasite-derived sequence, although the other replicate of this dilution did. Each of the two replicates generated for the other dilutions between 58,000 parasites/μL and 0.58 parasites/μL was positive using Ad_UPDx, suggestive of a technical error introduced during library preparation for that specific replicate only (Table 4, Fig. 3).

General comparison of nUPDx and Ad_UPDx

From DNA extraction to data analysis, Ad-UPDx took up to 5 days to complete, compared to close to 7 days for the nested UPDx method. Regarding the cost of each method, when considering the reagents for all digests, PCRs, ethanol washes, and bead cleanups during library preparation, Ad_UPDx was substantially cheaper at $11.24 per sample as compared to $40.40 for nUPDx. The cost of library preparation for nUPDx accounted for the bulk of the price difference ($29.79 per sample). These calculations assumed multiplexing of 80 samples within the same library. An overview of the two methods compared in this study is presented in Fig. 4.

Fig. 4.

Overview of the nUPDx method (Flaherty et al. 2021) and the newly described Ad_UPDx protocol. The comparison of the two approaches highlights the reduction of manipulation steps in Ad_UPDx, minimizing protocol complexity and leading to the reduction in turnaround time from 7 to 5 days and a cost decrease of analysis price per sample from $40.40 to $11.08

While both assays possessed a similar LOD, the proportion of total parasite-derived reads (for both paralogs) was higher for all replicates using Ad_UPDx compared to nUPDx (Table 4, Fig. 3). Furthermore, for clinical samples where the parasite load was expected to be low, e.g., the samples from patients with visceral leishmaniasis and chronic Chagas disease, Ad_UPDx generated higher percentages of parasite-derived reads, sometimes more than twice the proportions observed for nUPDx (Table 3, Fig. 2).

Discussion

We describe a new 18S metagenomic sequencing approach that simultaneously amplifies the 18S rRNA gene of eukaryotic pathogens while preparing the resultant amplicons for Miseq sequencing, similar to methods previously described for bacterial 16S metagenomics studies (Lee et al. 2019; Diaz-Torres et al. 2021). The introduction of an Illuminacompatible adapter and multiplexing index sequences to amplicons during PCR greatly reduced the complexity of preparing our pan-parasite TADS approach. The modifications to the original nUPDx method described here reduced turnaround times (from DNA extraction to generation of a result) by two days. Furthermore, the average cost per sample decreased from around $40 to $11, not considering the cost of human labor. Notably, relative to nUPDx, Ad_UPDx requires fewer “open-tube” steps and fewer reagents, reducing opportunities for contamination to occur.

The nUPDx assay included forward and reverse primers with priming sites approximately 1.5 kilobases (kb) apart, flanking the original ~ 200 bp UPDx amplicon. In addition to redesigning the UPDx primers for the generation of sequencing-ready amplicons, Ad-UPDx excluded this 1.5 kb amplification step by incorporating priming sites for the second PCR amplification during the first amplification. In this way, we retained the nested aspect of nUPDx while substantially reducing the size of the first-step amplicon.

As expected, Ad_UPDx and nUPDx were equally effective at detecting malaria species, Babesia species, kinetoplastids, and filarial nematodes in clinical blood samples. The LOD, as determined using samples spiked with cultured P. falciparum, was the same for the two methods. However, Ad_UPDx generated higher proportions of parasite-derived reads for samples with low amounts of parasites, such as the serially diluted samples with 0.58 parasites/μL and 5.8 parasites/μL, and some of the clinical samples. This indicates that Ad_UPDx is more robust compared to nUPDx for the detection of low parasitemia samples, likely because of a reduction in the size of the first PCR amplicon (0.2 kb compared to 1.5 kb), and other factors, such as the implementation of fewer steps for Ad_UPDx providing fewer opportunities for DNA loss.

As described above, one of two replicates containing the second-highest P. falciparum dilution (5800 parasites/μL) failed to produce a positive result using Ad_UPDx. Given that the other replicate at this dilution returned a strong positive result and that all subsequent dilutions down to 0.58 parasites/μL returned a positive (and with greater coverage than nUPDx for the matching dilutions), we attribute this to human error during assay preparation.

Both Ad-UPDx and nUPDx detected T. cruzi in acutely infected Chagas disease patients but not in chronically infected. The acute phase of Chagas disease lasts from 6 days to approximately 2 months (Barratt et al. 2010), as T. cruzi trypomastigotes migrate through the blood and lymph, remaining detectable via PCR. However, during the chronic phase of infection, T. cruzi becomes largely intracellular as it invades host cells to become amastigotes that possess a tropism for cardiac myocytes. During the intracellular chronic phase of Chagas disease, T. cruzi DNA exists at lower concentrations in the blood and lymph, making PCR detection difficult. The two samples from chronically infected patients included in this study had been confirmed positive for T. cruzi by real-time PCR assays targeting multicopy molecular targets commonly used for diagnostic assays (Qvarnstrom et al. 2012). The lower copy number of the 18S rRNA gene in T. cruzi, compared to these other targets (i.e., the kinetoplast minicircles and minichromosomal repeats) (Gonzalez et al. 1984; Sturm et al. 1989; Hernandez et al. 1993), may account for the failure of Ad_UPDx and nUPDx to detect T. cruzi DNA in the blood of chronically infected patients in this study.

While Ad_UPDx and nUPDx each differentiated between all human-infecting Plasmodium species, taxonomic resolution was achieved only at the genus or family level for Babesia spp., filarial nematodes, and Leishmania spp. The 18S rRNA genes are highly conserved across parasite taxa, making it difficult to distinguish all parasites at the species level using our short ~ 200 base pair UPDx amplicon. Future iterations of the UPDx method could incorporate additional markers with higher diversity to improve taxonomic resolution. However, the detection of parasite taxa at the family or genus level still provides valuable information that may subsequently guide diagnostic efforts and patient management, especially in cases of atypical disease presentation. Ad_UPDx might also be a good asset in other situations where a broadly reactive, agnostic detection method is warranted, for example, to screen donated blood samples prior to transfusion.

Disappointingly, mixed Plasmodium species infections were not detected in this study using nUPDx or Ad_UPDx, in contrast to the success in doing so in previous evaluations of UPDx (Flaherty et al. 2018, 2021). Mixed Plasmodium species infections are typically dominated by one species that exists at a much higher parasitemia than observed for the lesser species within a natural mixed infection (McKenzie and Bossert 1997). Thus, PCR amplification greatly favors the dominant species, resulting in an abundance of reads derived from that species, rendering the minor species either difficult to detect or un-detectable. Evaluation of samples from a variety of mixed species infections is currently underway to clarify the ability of Ad_UPDx to detect the minor species in these situations.

Conclusion

We describe a metagenomic sequencing approach combining nested amplification and restriction-enzyme-based enrichment of parasite-derived 18S rDNA PCR product, which results in amplicons that are immediately ready for Miseq sequencing. This procedure avoids the need for expensive and laborious library preparation, saving time and money, and reduces opportunities for contamination by decreasing the number of required steps and reagents. This improved 18S metagenomic sequencing method (named Ad_UPDx) detected and differentiated the same diversity of parasite taxa as earlier iterations of UPDx but performed more consistently at the lower detection limit of the assay and generated a greater proportion of parasite-derived reads generally.

Data availability

Raw reads analyzed in this study have been made publicly available by submission to NCBI Sequence Read Archive and can be accessed under BioProject accession number: PRJNA437674. BioSamples submitted to this BioProject that are relevant to the present study include the term “Ad_UPDx vs nUPDx” in their sample name..