(Extended Data) Amplicon deep sequencing of ama1 and mdr1 to track within-host P. falciparum diversity throughout treatment in a clinical drug trial

These extended data accompany the manuscript: Targeted Amplicon deep sequencing of ama1 and mdr1 to track within-host P. falciparum diversity throughout treatment in a clinical drug trial

Table S1: Concentration ratios and resulting parasitemia in artificial dna mixtures of P. falciparum Lab Isolates 3D7 and Dd2. This table presents the parasitemia for the artificial mixtures of P. falciparum lab isolates 3D7 and Dd2. Each mixture was prepared at varying ratios of 3D7 to Dd2, starting from equal proportions to a complete presence of only 3D7. The original concentration of each isolate was approximately 50,000 parasites per microliter (pf/μl), and the table displays the proportion of each strain in the mixture and the resulting total parasitemia concentration.

Table S2. List of PCR and deep sequencing primers. This table shows the list of forward and reverse primers used for deep sequencing. In boldface are the MID tags, while in the regular face are the forward primers

Table S3. The relative frequencies of each ama1 variant and the number of samples with each variant. The relative frequencies (%) of the 33 AMA1 variants in pre-and post-treatment samples (n = 330) are shown as a 33 amino acid sequence. The frequencies were calculated by dividing the number of reads of each microhaplotype by the total number of reads obtained per sample (116,187,131).

Table S4. Distribution of microhaplotypes among samples. This table shows the occurrence of microhaplotypes across all participants, both with monoclonal and multiclonal ama1 infections. It presents the ama1 clonality – monoclonal or multiclonal (column 1) - participant IDs (column 2), microhaplotype IDs (column 3), and the relative frequencies of these microhaplotypes across timepoints from 0 to 1008 hours (day 42) (column 3). Dashes represent time points where microhaplotypes were missing or were not detected.

Table S5. Distribution of rare microhaplotypes among samples. This table shows the occurrence of rare microhaplotypes in various samples. It presents participant IDs (column 1), microhaplotype IDs (column 2), and the relative frequencies of these microhaplotypes across time points from 0 to 1008 hours (day 42) (column 3). Samples containing rare microhaplotypes - specifically from PID10, PID32, PID38, PID40, PID49, PID60, PID63, and PID65 - are shown in orange, along with the corresponding rare microhaplotypes and their time points of occurrence. Furthermore, participants are categorised by shared microhaplotypes to indicate instances of rarity and commonality. Except for one microhaplotype unique to PID30, rare microhaplotypes were detected in several samples, frequently exceeding a 5% relative frequency. Dashes represent time points where microhaplotypes were missing or were not detected.

Table S6. The parasitemia levels associated with each ama1 microhaplotype per timepoint. This table shows the parasitemia for each ama1 microhaplotype per timepoint and each participant. “Patient ID” represents the patient ID, “AMA1 COI at 0h” represents the complexity of infection (COI) for each participant at baseline, based on ama1 while subsequent columns represent the parasitemia for each ama1 microhaplotype from timepoint 0h to 1008h. Parasitemia was back-calculated using the COI and total parasitemia for each time point. For time points with a COI > 1, parasitemia for the respective ama1 microhaplotypes are separated by commas, cells in red indicate timepoints without sequencing data (ND = not determined). In contrast, cells in grey indicate time points where microhaplotypes were detected below 10 parasites/μl, hence at risk of falling below the sampling limit.

Figure S1. Performance of AmpSeq in the sequencing controls. Six aliquots were prepared for each control set to ensure sufficient control data in case of PCR or sequencing failure. The median read depth in the lab controls was 5,658 (range 4,310 – 12,603) and 704 (291 – 1,676). The x-axis represents the aliquot identifier across the five mixtures, starting from 1 to 6, while the y-axis represents the proportions of each variant across all aliquots. For ama1 (A), two variants (3D7 and Dd2) were detected, whereas in mdr1 (B), two variants were detected YY, FY and NY following amplification of Dd2 Copy I, Dd2 Copy II and 3D7, respectively. For ama1, sequencing failed for aliquot 6 of control set 1, while for mdr1, sequencing failed for aliquot 2 and 6 of control set 3, aliquots 1 and 6 of control set 4 and aliquots 1 and 5 of control set 5. Under the mdr1 control set 4, the Dd2 copy II (86F, 184Y) was not identified, possibly due to having very low concentrations that were not picked up in this aliquot. Based on our control mixtures, the minimum variant frequency we could detect was 5%.

Figure S2. Heatmaps of the successfully PCR amplified and sequenced samples for ama1 (A) and mdr1 (B). The rows represent the study participants, while the columns represent time in hours. Successfully sequenced samples are shown in blue, those that failed PCR are shown in red and those that failed sequencing are in black. The timepoint “ Rec” represents unscheduled visits where a recurrent sample was collected. The unshaded areas with "-" are time points where samples were not collected. For each time point, the number of samples successfully sequenced (n Successful) is indicated in the last row of each panel. The table in panel C shows the groupings of samples based on parasitemia, high (> 5,000), moderate (100-5,000) and low (< 100 parasites per microlitre). Many samples collected between 0h-12h had high parasitemia, samples collected between 18h–30h had moderate parasitemia, while samples collected after 30h were primarily of low parasitemia.

Figure S3. The mean complexity of infection (COI) by AMA1 throughout treatment. The mean COI (red diamonds) appeared to be stable (between 1.5 - 2) from baseline (0h) up to 72h and thereafter fluctuated due to the small sample sizes (<5) in the post-treatment samples. The black dots represent the COI per sample.