A 336

For benchmarking ViralRecall on non-NCLDV viral sequences, we used a database of 879 non-NCLDV dsDNA genomes downloaded from NCBI. These genomes were selected because they are reference dsDNA viruses with genomes listed on the ICTV Virus Metadata Resource ([39 ]) and their genomes were available on NCBI RefSeq. The GVOG and Pfam normalization scores had been generated by using reference Caudovirales genomes in NCBI; therefore, we did not use genomes from this group that were present in the Virus Metadata Resource. Instead, we used a set of 336 jumbo bacteriophages (Caudovirales) that have been reported [40 (link)]. Additionally, we did not include any Lavidaviridae (virophage) in this set, because these viruses parasitize giant viruses, and, in some cases, may exchange genes with them [41 (link)]. To generate pseudocontigs for benchmarking, we used the gt-shredder command in genometools ([42 ]).
To benchmark ViralRecall on NCLDV sequences, we used a set of 1548 genomes in the NCLDV database described above. This included all genomes except those used in the construction of GVOGs, because those would not provide an unbiased assessment of the sensitivity of ViralRecall in detecting NCLDV sequences. For benchmarking purposes, ViralRecall was run with default parameters, with the only exception that the -c flag was used to generate mean contig-level scores.
For illustrative purposes we selected the following five NCLDV genomes from diverse families and provide the results generated by ViralRecall (shown in Figure 3): Acanthamoeba castellanii Medusavirus [43 (link)], Emiliania huxleyi virus 86 [7 (link)], Pithovirus sibericum [44 (link)], M. separata entomopoxvirus [45 (link)], and Hyperionvirus [46 (link)]. We also selected the genomes of four non-NCLDV dsDNA viruses for this purpose; we chose the jumbo bacteriophages with the highest mean score, lowest mean score, and longest length of those tested (FFC_PHAGE_43_1208, M01_PHAGE_56_67, and LP_PHAGE_CIR-CU-CL_32_18, respectively), as well as the human herpesvirus 3 strain Dumas [47 (link)]. Lastly, we also show the profiles for Yaravirus [48 (link)], a virus of A. castelanni with unclear evolutionary provenance, and the Sputnik virophage [41 (link)]. In all cases, the viral genomes shown here were not used in the construction of the GVOG database or for score normalization, thus they provide an unbiased assessment of ViralRecall results. For manual inspection of proteins encoded in contigs derived from suspected contamination, we performed homology searches against RefSeq v. 93 using BLASTP+ [49 (link)].
To illustrate how ViralRecall can be used to identify NCLDV signatures in eukaryotic genomes, we analyzed the Hydra vulgaris, Bigelowiella natans and Asterochloris glomerata genomes. Previous studies have already established NCLDV signatures in these genomes [19 (link),50 (link),51 (link)], and our results therefore provide independent verification.

The study protocol was approved by the Helsinki Health Centre Research Coordination Committee. Because the investigation involved only the review of patient records obtained during the course of medical care, no patient consent was required.
Data were collected from the dental records of 100 consecutive patients aged 40 years or older, who had been referred to the University Dental Clinic, Health Centre of Helsinki, for treatment of advanced chronic periodontitis during 2004–2005. Each patient was assigned an encrypted code linking the research data to the patient documents. Subjects with more than 14 teeth were included in the analysis. Patients using statins were compared with those who did not use statins.
We extracted the following data from the patient records: age, gender, reported current smoking, and use of statin medication. In addition, the records were checked for other medications, diabetes, rheumatoid diseases, and indicators of periodontal health.
A visible plaque index for index teeth [20 (link)] was extracted from the records.
For all teeth, we extracted recordings of six periodontal Probing Pocket Depth (PPD) values, measured to the nearest millimetre, from mesiobuccal, midbuccal, distobuccal, distolingual, midlingual, and mesiolingual surfaces using a WHO periodontal probe. [21 (link)-24 (link)]. Data on bleeding on probing was recorded for index teeth only and not included in the study. For a full dentition of 28 teeth, this yields 168 measurements. A gingival sulcus is considered physiological when PPD is less than 4 mm. The number of sites with moderate periodontal lesions (PPD at least 4 mm and less than 6 mm, modPPD) and advanced periodontal lesions (PPD 6 mm or deeper, advPPD) were recorded separately. For a dentition of 28 teeth, both of these indicators yield a maximum value of 6 × 28 = 168.
A novel index, Periodontal Inflammatory Burden Index or PIBI, was derived from the PPD values. The index is calculated by adding the number of periodontal sites indicating moderate periodontitis (N_modPPD) to the weighted number of periodontal sites indicating advanced periodontitis (N_advPPD).



If all measurement sites are advPPD in a 28-tooth dentition, PIBI can reach a maximum value of 336.

Neurons within the mapped visual cortex (~100–150 μm below the dura) were imaged at 4.22 Hz, using a Sutter MOM multiphoton microscope, in head-fixed awake mice restrained in a body tube. The Ti: sapphire laser (MaiTai HP: Newport SpectraPhysics; 940 nm) was routed to the microscope using table optics. The power was adjusted (20–40mW) using a rotating half-wave plate and a polarizing beam splitter to avoid signal saturation. A pair of galvanometric mirrors scan the laser beams to the back aperture of the objective (Nikon 16X 0.8 NA). The emission signal was collected through the same objective, passed through a short pass filter to block infrared wavelengths, and routed to a GaASP PMT after passing through a 540/50 bandpass filter. Image acquisition was controlled by Scanimage (Vidrio Technologies). The imaging field was a single Z frame of 336 × 336 um (256 × 256 pixels) consisting of 100 or more cells.

In vivo trials were performed using a previously described fish larvae method, ideal for bacterial virulence assays [40 (link)]. Briefly, good quality eggs of gilthead seabream, Sparus aurata, were collected and placed individually in 24-well plates (Cellstar, Greiner Bio-One, Kremsmünster, Austria). A total of 336 eggs in 14 plates were prepared and divided into three treatments (3 plates × 24 healthy eggs) and one control (5 plates × 24 healthy eggs). In the three treatments, the eggs were challenged with 10⁶ cells per mL (high bacterial dose) fresh overnight culture of the wild type V. anguillarum A023 and two lysogenized clones, respectively, whereas sterile SM buffer was added to the control plates. The plates were incubated at 18 °C and egg and larval mortality were monitored for seven consecutive days post infection (d.p.i.).

For morphological analysis, 30 histological slides per part of colon were examined, both in the control and in the diabetic group. The images for quantitative image analysis were obtained on an Olympus BX50 light microscope equipped with a Leica DFC 295 digital camera (Leica Micro-System, Reuil-Malmaison, France). The photomicrographs were taken at the magnification ×200.
Numerical areal density (N_A) of IPCs and ICC-IM, i.e., the average number of cells per mm² of the circular and longitudinal muscle layers, was determined with digital image analysis using the ImageJ software (National Institute of Health, Bethesda, MD, USA; http://imagej.nih.gov/ij/ (accessed on 18 December 2022)). N_A was determined by using the following formula: N_A = (10⁶ × N)/A, (N—number of the cells counted in muscle layer in the visual field, A—area of the muscle layer in mm²). We counted the number of these cells in the whole muscular layer in each slide. The cells were counted manually in order to avoid c-kit positive mast cells, which differ from ICC by their shape, granular content and location.
ICC-MP assessment was carried out by estimating the ICC-MP score (percentage encirclement of the ganglion by the processes of ICC-MPs), the semiquantitative method proposed by Den Braber-Ymker [30 (link)].
The analysis of the distribution of nerve elements was accomplished by determining the volume density of NF-M-positive fibers in the muscle layer. The volume density (Vv) is a relative variable, which shows how much overall space is occupied by the observed space in volume units. The Vv of nerve fibers and ganglia was determined by using ImageJ and a plugin of the software which inserted a grid system with 336 points (Vt). The number of points overlapping the nerve fibers and ganglia (Vf) within the colon muscle layer was counted. The Vv was determined using the following formula: Vv = Vf/Vt. The obtained results were multiplied by 100 and presented in percentages. In addition, the thickness of the circular and longitudinal muscle layer was determined using ImageJ software.